Public Types | Friends
ClpSimplex Class Reference

This solves LPs using the simplex method. More...

#include <ClpSimplex.hpp>

Inheritance diagram for ClpSimplex:
Inheritance graph
[legend]
Collaboration diagram for ClpSimplex:
Collaboration graph
[legend]

List of all members.

Public Types

enum  Status {
  isFree = 0x00, basic = 0x01, atUpperBound = 0x02, atLowerBound = 0x03,
  superBasic = 0x04, isFixed = 0x05
}
 enums for status of various sorts. More...
enum  FakeBound { noFake = 0x00, lowerFake = 0x01, upperFake = 0x02, bothFake = 0x03 }

Public Member Functions

Constructors and destructor and copy
 ClpSimplex (bool emptyMessages=false)
 Default constructor.
 ClpSimplex (const ClpSimplex &rhs, int scalingMode=-1)
 Copy constructor.
 ClpSimplex (const ClpModel &rhs, int scalingMode=-1)
 Copy constructor from model.
 ClpSimplex (const ClpModel *wholeModel, int numberRows, const int *whichRows, int numberColumns, const int *whichColumns, bool dropNames=true, bool dropIntegers=true, bool fixOthers=false)
 Subproblem constructor.
 ClpSimplex (const ClpSimplex *wholeModel, int numberRows, const int *whichRows, int numberColumns, const int *whichColumns, bool dropNames=true, bool dropIntegers=true, bool fixOthers=false)
 Subproblem constructor.
 ClpSimplex (ClpSimplex *wholeModel, int numberColumns, const int *whichColumns)
 This constructor modifies original ClpSimplex and stores original stuff in created ClpSimplex.
void originalModel (ClpSimplex *miniModel)
 This copies back stuff from miniModel and then deletes miniModel.
void setPersistenceFlag (int value)
 Array persistence flag If 0 then as now (delete/new) 1 then only do arrays if bigger needed 2 as 1 but give a bit extra if bigger needed.
void makeBaseModel ()
 Save a copy of model with certain state - normally without cuts.
void deleteBaseModel ()
 Switch off base model.
ClpSimplexbaseModel () const
 See if we have base model.
void setToBaseModel (ClpSimplex *model=NULL)
 Reset to base model (just size and arrays needed) If model NULL use internal copy.
ClpSimplexoperator= (const ClpSimplex &rhs)
 Assignment operator. This copies the data.
 ~ClpSimplex ()
 Destructor.
void loadProblem (const ClpMatrixBase &matrix, const double *collb, const double *colub, const double *obj, const double *rowlb, const double *rowub, const double *rowObjective=NULL)
 Loads a problem (the constraints on the rows are given by lower and upper bounds).
void loadProblem (const CoinPackedMatrix &matrix, const double *collb, const double *colub, const double *obj, const double *rowlb, const double *rowub, const double *rowObjective=NULL)
 Default constructor.
void loadProblem (const int numcols, const int numrows, const CoinBigIndex *start, const int *index, const double *value, const double *collb, const double *colub, const double *obj, const double *rowlb, const double *rowub, const double *rowObjective=NULL)
 Just like the other loadProblem() method except that the matrix is given in a standard column major ordered format (without gaps).
void loadProblem (const int numcols, const int numrows, const CoinBigIndex *start, const int *index, const double *value, const int *length, const double *collb, const double *colub, const double *obj, const double *rowlb, const double *rowub, const double *rowObjective=NULL)
 This one is for after presolve to save memory.
int loadProblem (CoinModel &modelObject, bool keepSolution=false)
 This loads a model from a coinModel object - returns number of errors.
int readMps (const char *filename, bool keepNames=false, bool ignoreErrors=false)
 Read an mps file from the given filename.
int readGMPL (const char *filename, const char *dataName, bool keepNames=false)
 Read GMPL files from the given filenames.
int readLp (const char *filename, const double epsilon=1e-5)
 Read file in LP format from file with name filename.
void borrowModel (ClpModel &otherModel)
 Borrow model.
void borrowModel (ClpSimplex &otherModel)
 Default constructor.
void passInEventHandler (const ClpEventHandler *eventHandler)
 Pass in Event handler (cloned and deleted at end)
void getbackSolution (const ClpSimplex &smallModel, const int *whichRow, const int *whichColumn)
 Puts solution back into small model.
int loadNonLinear (void *info, int &numberConstraints, ClpConstraint **&constraints)
 Load nonlinear part of problem from AMPL info Returns 0 if linear 1 if quadratic objective 2 if quadratic constraints 3 if nonlinear objective 4 if nonlinear constraints.
Functions most useful to user
int initialSolve (ClpSolve &options)
 General solve algorithm which can do presolve.
int initialSolve ()
 Default initial solve.
int initialDualSolve ()
 Dual initial solve.
int initialPrimalSolve ()
 Primal initial solve.
int initialBarrierSolve ()
 Barrier initial solve.
int initialBarrierNoCrossSolve ()
 Barrier initial solve, not to be followed by crossover.
int dual (int ifValuesPass=0, int startFinishOptions=0)
 Dual algorithm - see ClpSimplexDual.hpp for method.
int dualDebug (int ifValuesPass=0, int startFinishOptions=0)
 General solve algorithm which can do presolve.
int primal (int ifValuesPass=0, int startFinishOptions=0)
 Primal algorithm - see ClpSimplexPrimal.hpp for method.
int nonlinearSLP (int numberPasses, double deltaTolerance)
 Solves nonlinear problem using SLP - may be used as crash for other algorithms when number of iterations small.
int nonlinearSLP (int numberConstraints, ClpConstraint **constraints, int numberPasses, double deltaTolerance)
 Solves problem with nonlinear constraints using SLP - may be used as crash for other algorithms when number of iterations small.
int barrier (bool crossover=true)
 Solves using barrier (assumes you have good cholesky factor code).
int reducedGradient (int phase=0)
 Solves non-linear using reduced gradient.
int solve (CoinStructuredModel *model)
 Solve using structure of model and maybe in parallel.
int loadProblem (CoinStructuredModel &modelObject, bool originalOrder=true, bool keepSolution=false)
 This loads a model from a CoinStructuredModel object - returns number of errors.
int cleanup (int cleanupScaling)
 When scaling is on it is possible that the scaled problem is feasible but the unscaled is not.
int dualRanging (int numberCheck, const int *which, double *costIncrease, int *sequenceIncrease, double *costDecrease, int *sequenceDecrease, double *valueIncrease=NULL, double *valueDecrease=NULL)
 Dual ranging.
int primalRanging (int numberCheck, const int *which, double *valueIncrease, int *sequenceIncrease, double *valueDecrease, int *sequenceDecrease)
 Primal ranging.
int writeBasis (const char *filename, bool writeValues=false, int formatType=0) const
 Write the basis in MPS format to the specified file.
int readBasis (const char *filename)
 Read a basis from the given filename, returns -1 on file error, 0 if no values, 1 if values.
CoinWarmStartBasis * getBasis () const
 Returns a basis (to be deleted by user)
void setFactorization (ClpFactorization &factorization)
 Passes in factorization.
ClpFactorizationswapFactorization (ClpFactorization *factorization)
 General solve algorithm which can do presolve.
void copyFactorization (ClpFactorization &factorization)
 Copies in factorization to existing one.
int tightenPrimalBounds (double factor=0.0, int doTight=0, bool tightIntegers=false)
 Tightens primal bounds to make dual faster.
int crash (double gap, int pivot)
 Crash - at present just aimed at dual, returns.
void setDualRowPivotAlgorithm (ClpDualRowPivot &choice)
 Sets row pivot choice algorithm in dual.
void setPrimalColumnPivotAlgorithm (ClpPrimalColumnPivot &choice)
 Sets column pivot choice algorithm in primal.
int strongBranching (int numberVariables, const int *variables, double *newLower, double *newUpper, double **outputSolution, int *outputStatus, int *outputIterations, bool stopOnFirstInfeasible=true, bool alwaysFinish=false, int startFinishOptions=0)
 For strong branching.
int fathom (void *stuff)
 Fathom - 1 if solution.
int fathomMany (void *stuff)
 Do up to N deep - returns.
double doubleCheck ()
 Double checks OK.
int startFastDual2 (ClpNodeStuff *stuff)
 Starts Fast dual2.
int fastDual2 (ClpNodeStuff *stuff)
 Like Fast dual.
void stopFastDual2 (ClpNodeStuff *stuff)
 Stops Fast dual2.
ClpSimplexfastCrunch (ClpNodeStuff *stuff, int mode)
 Deals with crunch aspects mode 0 - in 1 - out with solution 2 - out without solution returns small model or NULL.
Needed for functionality of OsiSimplexInterface
int pivot ()
 Pivot in a variable and out a variable.
int primalPivotResult ()
 Pivot in a variable and choose an outgoing one.
int dualPivotResult ()
 Pivot out a variable and choose an incoing one.
int startup (int ifValuesPass, int startFinishOptions=0)
 Common bits of coding for dual and primal.
void finish (int startFinishOptions=0)
 Pivot in a variable and out a variable.
bool statusOfProblem (bool initial=false)
 Factorizes and returns true if optimal.
void defaultFactorizationFrequency ()
 If user left factorization frequency then compute.
most useful gets and sets
bool primalFeasible () const
 If problem is primal feasible.
bool dualFeasible () const
 If problem is dual feasible.
ClpFactorizationfactorization () const
 factorization
bool sparseFactorization () const
 Sparsity on or off.
void setSparseFactorization (bool value)
 If problem is primal feasible.
int factorizationFrequency () const
 Factorization frequency.
void setFactorizationFrequency (int value)
 If problem is primal feasible.
double dualBound () const
 Dual bound.
void setDualBound (double value)
 If problem is primal feasible.
double infeasibilityCost () const
 Infeasibility cost.
void setInfeasibilityCost (double value)
 If problem is primal feasible.
int perturbation () const
 Amount of print out: 0 - none 1 - just final 2 - just factorizations 3 - as 2 plus a bit more 4 - verbose above that 8,16,32 etc just for selective debug.
void setPerturbation (int value)
 If problem is primal feasible.
int algorithm () const
 Current (or last) algorithm.
void setAlgorithm (int value)
 Set algorithm.
bool isObjectiveLimitTestValid () const
 Return true if the objective limit test can be relied upon.
double sumDualInfeasibilities () const
 Sum of dual infeasibilities.
void setSumDualInfeasibilities (double value)
 If problem is primal feasible.
double sumOfRelaxedDualInfeasibilities () const
 Sum of relaxed dual infeasibilities.
void setSumOfRelaxedDualInfeasibilities (double value)
 If problem is primal feasible.
int numberDualInfeasibilities () const
 Number of dual infeasibilities.
void setNumberDualInfeasibilities (int value)
 If problem is primal feasible.
int numberDualInfeasibilitiesWithoutFree () const
 Number of dual infeasibilities (without free)
double sumPrimalInfeasibilities () const
 Sum of primal infeasibilities.
void setSumPrimalInfeasibilities (double value)
 If problem is primal feasible.
double sumOfRelaxedPrimalInfeasibilities () const
 Sum of relaxed primal infeasibilities.
void setSumOfRelaxedPrimalInfeasibilities (double value)
 If problem is primal feasible.
int numberPrimalInfeasibilities () const
 Number of primal infeasibilities.
void setNumberPrimalInfeasibilities (int value)
 If problem is primal feasible.
int saveModel (const char *fileName)
 Save model to file, returns 0 if success.
int restoreModel (const char *fileName)
 Restore model from file, returns 0 if success, deletes current model.
void checkSolution (int setToBounds=0)
 Just check solution (for external use) - sets sum of infeasibilities etc.
void checkSolutionInternal ()
 Just check solution (for internal use) - sets sum of infeasibilities etc.
CoinIndexedVector * rowArray (int index) const
 Useful row length arrays (0,1,2,3,4,5)
CoinIndexedVector * columnArray (int index) const
 Useful column length arrays (0,1,2,3,4,5)
double alphaAccuracy () const
 Initial value for alpha accuracy calculation (-1.0 off)
void setAlphaAccuracy (double value)
 If problem is primal feasible.
void setDisasterHandler (ClpDisasterHandler *handler)
 Objective value.
ClpDisasterHandlerdisasterHandler () const
 Get disaster handler.
double largeValue () const
 Large bound value (for complementarity etc)
void setLargeValue (double value)
 If problem is primal feasible.
double largestPrimalError () const
 Largest error on Ax-b.
double largestDualError () const
 Largest error on basic duals.
void setLargestPrimalError (double value)
 Largest error on Ax-b.
void setLargestDualError (double value)
 Largest error on basic duals.
double zeroTolerance () const
 Get zero tolerance.
void setZeroTolerance (double value)
 Set zero tolerance.
int * pivotVariable () const
 Basic variables pivoting on which rows.
bool automaticScaling () const
 If automatic scaling on.
void setAutomaticScaling (bool onOff)
 If problem is primal feasible.
double currentDualTolerance () const
 Current dual tolerance.
void setCurrentDualTolerance (double value)
 If problem is primal feasible.
double currentPrimalTolerance () const
 Current primal tolerance.
void setCurrentPrimalTolerance (double value)
 If problem is primal feasible.
int numberRefinements () const
 How many iterative refinements to do.
void setNumberRefinements (int value)
 If problem is primal feasible.
double alpha () const
 Alpha (pivot element) for use by classes e.g. steepestedge.
void setAlpha (double value)
 If problem is primal feasible.
double dualIn () const
 Reduced cost of last incoming for use by classes e.g. steepestedge.
int pivotRow () const
 Pivot Row for use by classes e.g. steepestedge.
void setPivotRow (int value)
 If problem is primal feasible.
double valueIncomingDual () const
 value of incoming variable (in Dual)
public methods
double * solutionRegion (int section) const
 Return row or column sections - not as much needed as it once was.
double * djRegion (int section) const
 Return row or column sections - not as much needed as it once was.
double * lowerRegion (int section) const
 Return row or column sections - not as much needed as it once was.
double * upperRegion (int section) const
 Return row or column sections - not as much needed as it once was.
double * costRegion (int section) const
 Return row or column sections - not as much needed as it once was.
double * solutionRegion () const
 Return region as single array.
double * djRegion () const
 Return row or column sections - not as much needed as it once was.
double * lowerRegion () const
 Return row or column sections - not as much needed as it once was.
double * upperRegion () const
 Return row or column sections - not as much needed as it once was.
double * costRegion () const
 Return row or column sections - not as much needed as it once was.
Status getStatus (int sequence) const
 Return row or column sections - not as much needed as it once was.
void setStatus (int sequence, Status newstatus)
 Return row or column sections - not as much needed as it once was.
bool startPermanentArrays ()
 Start or reset using maximumRows_ and Columns_ - true if change.
void setInitialDenseFactorization (bool onOff)
 Normally the first factorization does sparse coding because the factorization could be singular.
bool initialDenseFactorization () const
 Return row or column sections - not as much needed as it once was.
int sequenceIn () const
 Return sequence In or Out.
int sequenceOut () const
 Return row or column sections - not as much needed as it once was.
void setSequenceIn (int sequence)
 Set sequenceIn or Out.
void setSequenceOut (int sequence)
 Return row or column sections - not as much needed as it once was.
int directionIn () const
 Return direction In or Out.
int directionOut () const
 Return row or column sections - not as much needed as it once was.
void setDirectionIn (int direction)
 Set directionIn or Out.
void setDirectionOut (int direction)
 Return row or column sections - not as much needed as it once was.
double valueOut () const
 Value of Out variable.
void setValueOut (double value)
 Set value of out variable.
void setLowerOut (double value)
 Set lower of out variable.
void setUpperOut (double value)
 Set upper of out variable.
void setTheta (double value)
 Set theta of out variable.
int isColumn (int sequence) const
 Returns 1 if sequence indicates column.
int sequenceWithin (int sequence) const
 Returns sequence number within section.
double solution (int sequence)
 Return row or column values.
double & solutionAddress (int sequence)
 Return address of row or column values.
double reducedCost (int sequence)
 Return row or column sections - not as much needed as it once was.
double & reducedCostAddress (int sequence)
 Return row or column sections - not as much needed as it once was.
double lower (int sequence)
 Return row or column sections - not as much needed as it once was.
double & lowerAddress (int sequence)
 Return address of row or column lower bound.
double upper (int sequence)
 Return row or column sections - not as much needed as it once was.
double & upperAddress (int sequence)
 Return address of row or column upper bound.
double cost (int sequence)
 Return row or column sections - not as much needed as it once was.
double & costAddress (int sequence)
 Return address of row or column cost.
double originalLower (int iSequence) const
 Return original lower bound.
double originalUpper (int iSequence) const
 Return original lower bound.
double theta () const
 Theta (pivot change)
double bestPossibleImprovement () const
 Best possible improvement using djs (primal) or obj change by flipping bounds to make dual feasible (dual)
ClpNonLinearCostnonLinearCost () const
 Return pointer to details of costs.
int moreSpecialOptions () const
 Return more special options 1 bit - if presolve says infeasible in ClpSolve return 2 bit - if presolved problem infeasible return 4 bit - keep arrays like upper_ around 8 bit - if factorization kept can still declare optimal at once 16 bit - if checking replaceColumn accuracy before updating 32 bit - say optimal if primal feasible! 64 bit - give up easily in dual (and say infeasible) 128 bit - no objective, 0-1 and in B&B 256 bit - in primal from dual or vice versa 512 bit - alternative use of solveType_.
void setMoreSpecialOptions (int value)
 Set more special options 1 bit - if presolve says infeasible in ClpSolve return 2 bit - if presolved problem infeasible return 4 bit - keep arrays like upper_ around 8 bit - no free or superBasic variables 16 bit - if checking replaceColumn accuracy before updating 32 bit - say optimal if primal feasible! 64 bit - give up easily in dual (and say infeasible) 128 bit - no objective, 0-1 and in B&B 256 bit - in primal from dual or vice versa 512 bit - alternative use of solveType_.
status methods
void setFakeBound (int sequence, FakeBound fakeBound)
 To flag a variable (not inline to allow for column generation)
FakeBound getFakeBound (int sequence) const
 To flag a variable (not inline to allow for column generation)
void setRowStatus (int sequence, Status newstatus)
 To flag a variable (not inline to allow for column generation)
Status getRowStatus (int sequence) const
 To flag a variable (not inline to allow for column generation)
void setColumnStatus (int sequence, Status newstatus)
 To flag a variable (not inline to allow for column generation)
Status getColumnStatus (int sequence) const
 To flag a variable (not inline to allow for column generation)
void setPivoted (int sequence)
 To flag a variable (not inline to allow for column generation)
void clearPivoted (int sequence)
 To flag a variable (not inline to allow for column generation)
bool pivoted (int sequence) const
 To flag a variable (not inline to allow for column generation)
void setFlagged (int sequence)
 To flag a variable (not inline to allow for column generation)
void clearFlagged (int sequence)
 To flag a variable (not inline to allow for column generation)
bool flagged (int sequence) const
 To flag a variable (not inline to allow for column generation)
void setActive (int iRow)
 To say row active in primal pivot row choice.
void clearActive (int iRow)
 To flag a variable (not inline to allow for column generation)
bool active (int iRow) const
 To flag a variable (not inline to allow for column generation)
void createStatus ()
 Set up status array (can be used by OsiClp).
void allSlackBasis (bool resetSolution=false)
 Sets up all slack basis and resets solution to as it was after initial load or readMps.
int lastBadIteration () const
 So we know when to be cautious.
int progressFlag () const
 Progress flag - at present 0 bit says artificials out.
void forceFactorization (int value)
 Force re-factorization early.
double rawObjectiveValue () const
 Raw objective value (so always minimize in primal)
void computeObjectiveValue (bool useWorkingSolution=false)
 Compute objective value from solution and put in objectiveValue_.
double computeInternalObjectiveValue ()
 Compute minimization objective value from internal solution without perturbation.
int numberExtraRows () const
 Number of extra rows.
int maximumBasic () const
 Maximum number of basic variables - can be more than number of rows if GUB.
int baseIteration () const
 Iteration when we entered dual or primal.
void generateCpp (FILE *fp, bool defaultFactor=false)
 Create C++ lines to get to current state.
ClpFactorizationgetEmptyFactorization ()
 Gets clean and emptyish factorization.
void setEmptyFactorization ()
 May delete or may make clean and emptyish factorization.
void moveInfo (const ClpSimplex &rhs, bool justStatus=false)
 Move status and solution across.
Basis handling
void getBInvARow (int row, double *z, double *slack=NULL)
 Get a row of the tableau (slack part in slack if not NULL)
void getBInvRow (int row, double *z)
 Get a row of the basis inverse.
void getBInvACol (int col, double *vec)
 Get a column of the tableau.
void getBInvCol (int col, double *vec)
 Get a column of the basis inverse.
void getBasics (int *index)
 Get basic indices (order of indices corresponds to the order of elements in a vector retured by getBInvACol() and getBInvCol()).
Changing bounds on variables and constraints
void setObjectiveCoefficient (int elementIndex, double elementValue)
 Set an objective function coefficient.
void setObjCoeff (int elementIndex, double elementValue)
 Set an objective function coefficient.
void setColumnLower (int elementIndex, double elementValue)
 Set a single column lower bound
Use -DBL_MAX for -infinity.
void setColumnUpper (int elementIndex, double elementValue)
 Set a single column upper bound
Use DBL_MAX for infinity.
void setColumnBounds (int elementIndex, double lower, double upper)
 Set a single column lower and upper bound.
void setColumnSetBounds (const int *indexFirst, const int *indexLast, const double *boundList)
 Set the bounds on a number of columns simultaneously
The default implementation just invokes setColLower() and setColUpper() over and over again.
void setColLower (int elementIndex, double elementValue)
 Set a single column lower bound
Use -DBL_MAX for -infinity.
void setColUpper (int elementIndex, double elementValue)
 Set a single column upper bound
Use DBL_MAX for infinity.
void setColBounds (int elementIndex, double newlower, double newupper)
 Set a single column lower and upper bound.
void setColSetBounds (const int *indexFirst, const int *indexLast, const double *boundList)
 Set the bounds on a number of columns simultaneously

void setRowLower (int elementIndex, double elementValue)
 Set a single row lower bound
Use -DBL_MAX for -infinity.
void setRowUpper (int elementIndex, double elementValue)
 Set a single row upper bound
Use DBL_MAX for infinity.
void setRowBounds (int elementIndex, double lower, double upper)
 Set a single row lower and upper bound.
void setRowSetBounds (const int *indexFirst, const int *indexLast, const double *boundList)
 Set the bounds on a number of rows simultaneously

void resize (int newNumberRows, int newNumberColumns)
 Resizes rim part of model.

Protected Member Functions

protected methods
int gutsOfSolution (double *givenDuals, const double *givenPrimals, bool valuesPass=false)
 May change basis and then returns number changed.
void gutsOfDelete (int type)
 Does most of deletion (0 = all, 1 = most, 2 most + factorization)
void gutsOfCopy (const ClpSimplex &rhs)
 Does most of copying.
bool createRim (int what, bool makeRowCopy=false, int startFinishOptions=0)
 puts in format I like (rowLower,rowUpper) also see StandardMatrix 1 bit does rows (now and columns), (2 bit does column bounds), 4 bit does objective(s).
void createRim1 (bool initial)
 Does rows and columns.
void createRim4 (bool initial)
 Does objective.
void createRim5 (bool initial)
 Does rows and columns and objective.
void deleteRim (int getRidOfFactorizationData=2)
 releases above arrays and does solution scaling out.
bool sanityCheck ()
 Sanity check on input rim data (after scaling) - returns true if okay.

Friends

void ClpSimplexUnitTest (const std::string &mpsDir)
 A function that tests the methods in the ClpSimplex class.

Functions less likely to be useful to casual user

int getSolution (const double *rowActivities, const double *columnActivities)
 Given an existing factorization computes and checks primal and dual solutions.
int getSolution ()
 Given an existing factorization computes and checks primal and dual solutions.
int createPiecewiseLinearCosts (const int *starts, const double *lower, const double *gradient)
 Constructs a non linear cost from list of non-linearities (columns only) First lower of each column is taken as real lower Last lower is taken as real upper and cost ignored.
ClpDualRowPivotdualRowPivot () const
 dual row pivot choice
ClpPrimalColumnPivotprimalColumnPivot () const
 primal column pivot choice
bool goodAccuracy () const
 Returns true if model looks OK.
void returnModel (ClpSimplex &otherModel)
 Return model - updates any scalars.
int internalFactorize (int solveType)
 Factorizes using current basis.
ClpDataSave saveData ()
 Save data.
void restoreData (ClpDataSave saved)
 Restore data.
void cleanStatus ()
 Clean up status.
int factorize ()
 Factorizes using current basis. For external use.
void computeDuals (double *givenDjs)
 Computes duals from scratch.
void computePrimals (const double *rowActivities, const double *columnActivities)
 Computes primals from scratch.
void add (double *array, int column, double multiplier) const
 Adds multiple of a column into an array.
void unpack (CoinIndexedVector *rowArray) const
 Unpacks one column of the matrix into indexed array Uses sequenceIn_ Also applies scaling if needed.
void unpack (CoinIndexedVector *rowArray, int sequence) const
 Unpacks one column of the matrix into indexed array Slack if sequence>= numberColumns Also applies scaling if needed.
void unpackPacked (CoinIndexedVector *rowArray)
 Unpacks one column of the matrix into indexed array as packed vector Uses sequenceIn_ Also applies scaling if needed.
void unpackPacked (CoinIndexedVector *rowArray, int sequence)
 Unpacks one column of the matrix into indexed array as packed vector Slack if sequence>= numberColumns Also applies scaling if needed.
void setValuesPassAction (double incomingInfeasibility, double allowedInfeasibility)
 For advanced use.
int housekeeping (double objectiveChange)
 This does basis housekeeping and does values for in/out variables.
void checkPrimalSolution (const double *rowActivities=NULL, const double *columnActivies=NULL)
 This sets largest infeasibility and most infeasible and sum and number of infeasibilities (Primal)
void checkDualSolution ()
 This sets largest infeasibility and most infeasible and sum and number of infeasibilities (Dual)
void checkBothSolutions ()
 This sets sum and number of infeasibilities (Dual and Primal)
double scaleObjective (double value)
 If input negative scales objective so maximum <= -value and returns scale factor used.
int solveDW (CoinStructuredModel *model)
 Solve using Dantzig-Wolfe decomposition and maybe in parallel.
int solveBenders (CoinStructuredModel *model)
 Solve using Benders decomposition and maybe in parallel.

data. Many arrays have a row part and a column part.

There is a single array with both - columns then rows and then normally two arrays pointing to rows and columns.

The single array is the owner of memory

double bestPossibleImprovement_
 Best possible improvement using djs (primal) or obj change by flipping bounds to make dual feasible (dual)
double zeroTolerance_
 Zero tolerance.
int columnPrimalSequence_
 Sequence of worst (-1 if feasible)
int rowPrimalSequence_
 Sequence of worst (-1 if feasible)
double bestObjectiveValue_
 "Best" objective value
int moreSpecialOptions_
 More special options - see set for details.
int baseIteration_
 Iteration when we entered dual or primal.
double primalToleranceToGetOptimal_
 Primal tolerance needed to make dual feasible (<largeTolerance)
double largeValue_
 Large bound value (for complementarity etc)
double largestPrimalError_
 Largest error on Ax-b.
double largestDualError_
 Largest error on basic duals.
double alphaAccuracy_
 For computing whether to re-factorize.
double dualBound_
 Dual bound.
double alpha_
 Alpha (pivot element)
double theta_
 Theta (pivot change)
double lowerIn_
 Lower Bound on In variable.
double valueIn_
 Value of In variable.
double upperIn_
 Upper Bound on In variable.
double dualIn_
 Reduced cost of In variable.
double lowerOut_
 Lower Bound on Out variable.
double valueOut_
 Value of Out variable.
double upperOut_
 Upper Bound on Out variable.
double dualOut_
 Infeasibility (dual) or ? (primal) of Out variable.
double dualTolerance_
 Current dual tolerance for algorithm.
double primalTolerance_
 Current primal tolerance for algorithm.
double sumDualInfeasibilities_
 Sum of dual infeasibilities.
double sumPrimalInfeasibilities_
 Sum of primal infeasibilities.
double infeasibilityCost_
 Weight assigned to being infeasible in primal.
double sumOfRelaxedDualInfeasibilities_
 Sum of Dual infeasibilities using tolerance based on error in duals.
double sumOfRelaxedPrimalInfeasibilities_
 Sum of Primal infeasibilities using tolerance based on error in primals.
double acceptablePivot_
 Acceptable pivot value just after factorization.
double * lower_
 Working copy of lower bounds (Owner of arrays below)
double * rowLowerWork_
 Row lower bounds - working copy.
double * columnLowerWork_
 Column lower bounds - working copy.
double * upper_
 Working copy of upper bounds (Owner of arrays below)
double * rowUpperWork_
 Row upper bounds - working copy.
double * columnUpperWork_
 Column upper bounds - working copy.
double * cost_
 Working copy of objective (Owner of arrays below)
double * rowObjectiveWork_
 Row objective - working copy.
double * objectiveWork_
 Column objective - working copy.
CoinIndexedVector * rowArray_ [6]
 Useful row length arrays.
CoinIndexedVector * columnArray_ [6]
 Useful column length arrays.
int sequenceIn_
 Sequence of In variable.
int directionIn_
 Direction of In, 1 going up, -1 going down, 0 not a clude.
int sequenceOut_
 Sequence of Out variable.
int directionOut_
 Direction of Out, 1 to upper bound, -1 to lower bound, 0 - superbasic.
int pivotRow_
 Pivot Row.
int lastGoodIteration_
 Last good iteration (immediately after a re-factorization)
double * dj_
 Working copy of reduced costs (Owner of arrays below)
double * rowReducedCost_
 Reduced costs of slacks not same as duals (or - duals)
double * reducedCostWork_
 Possible scaled reduced costs.
double * solution_
 Working copy of primal solution (Owner of arrays below)
double * rowActivityWork_
 Row activities - working copy.
double * columnActivityWork_
 Column activities - working copy.
int numberDualInfeasibilities_
 Number of dual infeasibilities.
int numberDualInfeasibilitiesWithoutFree_
 Number of dual infeasibilities (without free)
int numberPrimalInfeasibilities_
 Number of primal infeasibilities.
int numberRefinements_
 How many iterative refinements to do.
ClpDualRowPivotdualRowPivot_
 dual row pivot choice
ClpPrimalColumnPivotprimalColumnPivot_
 primal column pivot choice
int * pivotVariable_
 Basic variables pivoting on which rows.
ClpFactorizationfactorization_
 factorization
double * savedSolution_
 Saved version of solution.
int numberTimesOptimal_
 Number of times code has tentatively thought optimal.
ClpDisasterHandlerdisasterArea_
 Disaster handler.
int changeMade_
 If change has been made (first attempt at stopping looping)
int algorithm_
 Algorithm >0 == Primal, <0 == Dual.
int forceFactorization_
 Now for some reliability aids This forces re-factorization early.
int perturbation_
 Perturbation:
unsigned char * saveStatus_
 Saved status regions.
ClpNonLinearCostnonLinearCost_
 Very wasteful way of dealing with infeasibilities in primal.
int lastBadIteration_
 So we know when to be cautious.
int lastFlaggedIteration_
 So we know when to open up again.
int numberFake_
 Can be used for count of fake bounds (dual) or fake costs (primal)
int numberChanged_
 Can be used for count of changed costs (dual) or changed bounds (primal)
int progressFlag_
 Progress flag - at present 0 bit says artificials out, 1 free in.
int firstFree_
 First free/super-basic variable (-1 if none)
int numberExtraRows_
 Number of extra rows.
int maximumBasic_
 Maximum number of basic variables - can be more than number of rows if GUB.
int dontFactorizePivots_
 If may skip final factorize then allow up to this pivots (default 20)
double incomingInfeasibility_
 For advanced use.
double allowedInfeasibility_
 Best possible improvement using djs (primal) or obj change by flipping bounds to make dual feasible (dual)
int automaticScale_
 Automatic scaling of objective and rhs and bounds.
int maximumPerturbationSize_
 Maximum perturbation array size (take out when code rewritten)
double * perturbationArray_
 Perturbation array (maximumPerturbationSize_)
ClpSimplexbaseModel_
 A copy of model with certain state - normally without cuts.
ClpSimplexProgress progress_
 For dealing with all issues of cycling etc.
int spareIntArray_ [4]
 Spare int array for passing information [0]!=0 switches on.
double spareDoubleArray_ [4]
 Spare double array for passing information [0]!=0 switches on.
class OsiClpSolverInterface
 Allow OsiClp certain perks.

Detailed Description

This solves LPs using the simplex method.

It inherits from ClpModel and all its arrays are created at algorithm time. Originally I tried to work with model arrays but for simplicity of coding I changed to single arrays with structural variables then row variables. Some coding is still based on old style and needs cleaning up.

For a description of algorithms:

for dual see ClpSimplexDual.hpp and at top of ClpSimplexDual.cpp for primal see ClpSimplexPrimal.hpp and at top of ClpSimplexPrimal.cpp

There is an algorithm data member. + for primal variations and - for dual variations

Definition at line 49 of file ClpSimplex.hpp.


Member Enumeration Documentation

enums for status of various sorts.

First 4 match CoinWarmStartBasis, isFixed means fixed at lower bound and out of basis

Enumerator:
isFree 
basic 
atUpperBound 
atLowerBound 
superBasic 
isFixed 

Definition at line 57 of file ClpSimplex.hpp.

Enumerator:
noFake 
lowerFake 
upperFake 
bothFake 

Definition at line 66 of file ClpSimplex.hpp.


Constructor & Destructor Documentation

ClpSimplex::ClpSimplex ( bool  emptyMessages = false)

Default constructor.

ClpSimplex::ClpSimplex ( const ClpSimplex rhs,
int  scalingMode = -1 
)

Copy constructor.

May scale depending on mode -1 leave mode as is 0 -off, 1 equilibrium, 2 geometric, 3, auto, 4 dynamic(later)

ClpSimplex::ClpSimplex ( const ClpModel rhs,
int  scalingMode = -1 
)

Copy constructor from model.

May scale depending on mode -1 leave mode as is 0 -off, 1 equilibrium, 2 geometric, 3, auto, 4 dynamic(later)

ClpSimplex::ClpSimplex ( const ClpModel wholeModel,
int  numberRows,
const int *  whichRows,
int  numberColumns,
const int *  whichColumns,
bool  dropNames = true,
bool  dropIntegers = true,
bool  fixOthers = false 
)

Subproblem constructor.

A subset of whole model is created from the row and column lists given. The new order is given by list order and duplicates are allowed. Name and integer information can be dropped Can optionally modify rhs to take into account variables NOT in list in this case duplicates are not allowed (also see getbackSolution)

ClpSimplex::ClpSimplex ( const ClpSimplex wholeModel,
int  numberRows,
const int *  whichRows,
int  numberColumns,
const int *  whichColumns,
bool  dropNames = true,
bool  dropIntegers = true,
bool  fixOthers = false 
)

Subproblem constructor.

A subset of whole model is created from the row and column lists given. The new order is given by list order and duplicates are allowed. Name and integer information can be dropped Can optionally modify rhs to take into account variables NOT in list in this case duplicates are not allowed (also see getbackSolution)

ClpSimplex::ClpSimplex ( ClpSimplex wholeModel,
int  numberColumns,
const int *  whichColumns 
)

This constructor modifies original ClpSimplex and stores original stuff in created ClpSimplex.

It is only to be used in conjunction with originalModel

Destructor.


Member Function Documentation

void ClpSimplex::originalModel ( ClpSimplex miniModel)

This copies back stuff from miniModel and then deletes miniModel.

Only to be used with mini constructor

void ClpSimplex::setPersistenceFlag ( int  value)

Array persistence flag If 0 then as now (delete/new) 1 then only do arrays if bigger needed 2 as 1 but give a bit extra if bigger needed.

Save a copy of model with certain state - normally without cuts.

Switch off base model.

ClpSimplex* ClpSimplex::baseModel ( ) const [inline]

See if we have base model.

Definition at line 129 of file ClpSimplex.hpp.

void ClpSimplex::setToBaseModel ( ClpSimplex model = NULL)

Reset to base model (just size and arrays needed) If model NULL use internal copy.

ClpSimplex& ClpSimplex::operator= ( const ClpSimplex rhs)

Assignment operator. This copies the data.

void ClpSimplex::loadProblem ( const ClpMatrixBase matrix,
const double *  collb,
const double *  colub,
const double *  obj,
const double *  rowlb,
const double *  rowub,
const double *  rowObjective = NULL 
)

Loads a problem (the constraints on the rows are given by lower and upper bounds).

If a pointer is 0 then the following values are the default:

  • colub: all columns have upper bound infinity
  • collb: all columns have lower bound 0
  • rowub: all rows have upper bound infinity
  • rowlb: all rows have lower bound -infinity
  • obj: all variables have 0 objective coefficient

Reimplemented from ClpModel.

void ClpSimplex::loadProblem ( const CoinPackedMatrix &  matrix,
const double *  collb,
const double *  colub,
const double *  obj,
const double *  rowlb,
const double *  rowub,
const double *  rowObjective = NULL 
)

Default constructor.

Reimplemented from ClpModel.

void ClpSimplex::loadProblem ( const int  numcols,
const int  numrows,
const CoinBigIndex *  start,
const int *  index,
const double *  value,
const double *  collb,
const double *  colub,
const double *  obj,
const double *  rowlb,
const double *  rowub,
const double *  rowObjective = NULL 
)

Just like the other loadProblem() method except that the matrix is given in a standard column major ordered format (without gaps).

Reimplemented from ClpModel.

void ClpSimplex::loadProblem ( const int  numcols,
const int  numrows,
const CoinBigIndex *  start,
const int *  index,
const double *  value,
const int *  length,
const double *  collb,
const double *  colub,
const double *  obj,
const double *  rowlb,
const double *  rowub,
const double *  rowObjective = NULL 
)

This one is for after presolve to save memory.

Reimplemented from ClpModel.

int ClpSimplex::loadProblem ( CoinModel &  modelObject,
bool  keepSolution = false 
)

This loads a model from a coinModel object - returns number of errors.

If keepSolution true and size is same as current then keeps current status and solution

Reimplemented from ClpModel.

int ClpSimplex::readMps ( const char *  filename,
bool  keepNames = false,
bool  ignoreErrors = false 
)

Read an mps file from the given filename.

Reimplemented from ClpModel.

int ClpSimplex::readGMPL ( const char *  filename,
const char *  dataName,
bool  keepNames = false 
)

Read GMPL files from the given filenames.

Reimplemented from ClpModel.

int ClpSimplex::readLp ( const char *  filename,
const double  epsilon = 1e-5 
)

Read file in LP format from file with name filename.

See class CoinLpIO for description of this format.

void ClpSimplex::borrowModel ( ClpModel otherModel)

Borrow model.

This is so we dont have to copy large amounts of data around. It assumes a derived class wants to overwrite an empty model with a real one - while it does an algorithm. This is same as ClpModel one, but sets scaling on etc.

Reimplemented from ClpModel.

void ClpSimplex::borrowModel ( ClpSimplex otherModel)

Default constructor.

void ClpSimplex::passInEventHandler ( const ClpEventHandler eventHandler)

Pass in Event handler (cloned and deleted at end)

Reimplemented from ClpModel.

void ClpSimplex::getbackSolution ( const ClpSimplex smallModel,
const int *  whichRow,
const int *  whichColumn 
)

Puts solution back into small model.

int ClpSimplex::loadNonLinear ( void *  info,
int &  numberConstraints,
ClpConstraint **&  constraints 
)

Load nonlinear part of problem from AMPL info Returns 0 if linear 1 if quadratic objective 2 if quadratic constraints 3 if nonlinear objective 4 if nonlinear constraints.

-1 on failure

int ClpSimplex::initialSolve ( ClpSolve options)

General solve algorithm which can do presolve.

See ClpSolve.hpp for options

Default initial solve.

Dual initial solve.

Primal initial solve.

Barrier initial solve.

Barrier initial solve, not to be followed by crossover.

int ClpSimplex::dual ( int  ifValuesPass = 0,
int  startFinishOptions = 0 
)

Dual algorithm - see ClpSimplexDual.hpp for method.

ifValuesPass==2 just does values pass and then stops.

startFinishOptions - bits 1 - do not delete work areas and factorization at end 2 - use old factorization if same number of rows 4 - skip as much initialization of work areas as possible (based on whatsChanged in clpmodel.hpp) ** work in progress maybe other bits later

Reimplemented in ClpSimplexDual.

int ClpSimplex::dualDebug ( int  ifValuesPass = 0,
int  startFinishOptions = 0 
)

General solve algorithm which can do presolve.

See ClpSolve.hpp for options

int ClpSimplex::primal ( int  ifValuesPass = 0,
int  startFinishOptions = 0 
)

Primal algorithm - see ClpSimplexPrimal.hpp for method.

ifValuesPass==2 just does values pass and then stops.

startFinishOptions - bits 1 - do not delete work areas and factorization at end 2 - use old factorization if same number of rows 4 - skip as much initialization of work areas as possible (based on whatsChanged in clpmodel.hpp) ** work in progress maybe other bits later

Reimplemented in ClpSimplexPrimal.

int ClpSimplex::nonlinearSLP ( int  numberPasses,
double  deltaTolerance 
)

Solves nonlinear problem using SLP - may be used as crash for other algorithms when number of iterations small.

Also exits if all problematical variables are changing less than deltaTolerance

int ClpSimplex::nonlinearSLP ( int  numberConstraints,
ClpConstraint **  constraints,
int  numberPasses,
double  deltaTolerance 
)

Solves problem with nonlinear constraints using SLP - may be used as crash for other algorithms when number of iterations small.

Also exits if all problematical variables are changing less than deltaTolerance

int ClpSimplex::barrier ( bool  crossover = true)

Solves using barrier (assumes you have good cholesky factor code).

Does crossover to simplex if asked

int ClpSimplex::reducedGradient ( int  phase = 0)

Solves non-linear using reduced gradient.

Phase = 0 get feasible, =1 use solution

int ClpSimplex::solve ( CoinStructuredModel *  model)

Solve using structure of model and maybe in parallel.

int ClpSimplex::loadProblem ( CoinStructuredModel &  modelObject,
bool  originalOrder = true,
bool  keepSolution = false 
)

This loads a model from a CoinStructuredModel object - returns number of errors.

If originalOrder then keep to order stored in blocks, otherwise first column/rows correspond to first block - etc. If keepSolution true and size is same as current then keeps current status and solution

int ClpSimplex::cleanup ( int  cleanupScaling)

When scaling is on it is possible that the scaled problem is feasible but the unscaled is not.

Clp returns a secondary status code to that effect. This option allows for a cleanup. If you use it I would suggest 1. This only affects actions when scaled optimal 0 - no action 1 - clean up using dual if primal infeasibility 2 - clean up using dual if dual infeasibility 3 - clean up using dual if primal or dual infeasibility 11,12,13 - as 1,2,3 but use primal

return code as dual/primal

int ClpSimplex::dualRanging ( int  numberCheck,
const int *  which,
double *  costIncrease,
int *  sequenceIncrease,
double *  costDecrease,
int *  sequenceDecrease,
double *  valueIncrease = NULL,
double *  valueDecrease = NULL 
)

Dual ranging.

This computes increase/decrease in cost for each given variable and corresponding sequence numbers which would change basis. Sequence numbers are 0..numberColumns and numberColumns.. for artificials/slacks. For non-basic variables the information is trivial to compute and the change in cost is just minus the reduced cost and the sequence number will be that of the non-basic variables. For basic variables a ratio test is between the reduced costs for non-basic variables and the row of the tableau corresponding to the basic variable. The increase/decrease value is always >= 0.0

Up to user to provide correct length arrays where each array is of length numberCheck. which contains list of variables for which information is desired. All other arrays will be filled in by function. If fifth entry in which is variable 7 then fifth entry in output arrays will be information for variable 7.

If valueIncrease/Decrease not NULL (both must be NULL or both non NULL) then these are filled with the value of variable if such a change in cost were made (the existing bounds are ignored)

Returns non-zero if infeasible unbounded etc

Reimplemented in ClpSimplexOther.

int ClpSimplex::primalRanging ( int  numberCheck,
const int *  which,
double *  valueIncrease,
int *  sequenceIncrease,
double *  valueDecrease,
int *  sequenceDecrease 
)

Primal ranging.

This computes increase/decrease in value for each given variable and corresponding sequence numbers which would change basis. Sequence numbers are 0..numberColumns and numberColumns.. for artificials/slacks. This should only be used for non-basic variabls as otherwise information is pretty useless For basic variables the sequence number will be that of the basic variables.

Up to user to provide correct length arrays where each array is of length numberCheck. which contains list of variables for which information is desired. All other arrays will be filled in by function. If fifth entry in which is variable 7 then fifth entry in output arrays will be information for variable 7.

Returns non-zero if infeasible unbounded etc

Reimplemented in ClpSimplexOther.

int ClpSimplex::writeBasis ( const char *  filename,
bool  writeValues = false,
int  formatType = 0 
) const

Write the basis in MPS format to the specified file.

If writeValues true writes values of structurals (and adds VALUES to end of NAME card)

Row and column names may be null. formatType is

  • 0 - normal
  • 1 - extra accuracy
  • 2 - IEEE hex (later)

Returns non-zero on I/O error

Reimplemented in ClpSimplexOther.

int ClpSimplex::readBasis ( const char *  filename)

Read a basis from the given filename, returns -1 on file error, 0 if no values, 1 if values.

Reimplemented in ClpSimplexOther.

CoinWarmStartBasis* ClpSimplex::getBasis ( ) const

Returns a basis (to be deleted by user)

Passes in factorization.

General solve algorithm which can do presolve.

See ClpSolve.hpp for options

Copies in factorization to existing one.

int ClpSimplex::tightenPrimalBounds ( double  factor = 0.0,
int  doTight = 0,
bool  tightIntegers = false 
)

Tightens primal bounds to make dual faster.

Unless fixed or doTight>10, bounds are slightly looser than they could be. This is to make dual go faster and is probably not needed with a presolve. Returns non-zero if problem infeasible.

Fudge for branch and bound - put bounds on columns of factor * largest value (at continuous) - should improve stability in branch and bound on infeasible branches (0.0 is off)

int ClpSimplex::crash ( double  gap,
int  pivot 
)

Crash - at present just aimed at dual, returns.

-2 if dual preferred and crash basis created -1 if dual preferred and all slack basis preferred 0 if basis going in was not all slack 1 if primal preferred and all slack basis preferred 2 if primal preferred and crash basis created.

if gap between bounds <="gap" variables can be flipped ( If pivot -1 then can be made super basic!)

If "pivot" is -1 No pivoting - always primal 0 No pivoting (so will just be choice of algorithm) 1 Simple pivoting e.g. gub 2 Mini iterations

Sets row pivot choice algorithm in dual.

Sets column pivot choice algorithm in primal.

int ClpSimplex::strongBranching ( int  numberVariables,
const int *  variables,
double *  newLower,
double *  newUpper,
double **  outputSolution,
int *  outputStatus,
int *  outputIterations,
bool  stopOnFirstInfeasible = true,
bool  alwaysFinish = false,
int  startFinishOptions = 0 
)

For strong branching.

On input lower and upper are new bounds while on output they are change in objective function values (>1.0e50 infeasible). Return code is 0 if nothing interesting, -1 if infeasible both ways and +1 if infeasible one way (check values to see which one(s)) Solutions are filled in as well - even down, odd up - also status and number of iterations

Reimplemented in ClpSimplexDual.

int ClpSimplex::fathom ( void *  stuff)

Fathom - 1 if solution.

int ClpSimplex::fathomMany ( void *  stuff)

Do up to N deep - returns.

-1 - no solution nNodes_ valid nodes >= if solution and that node gives solution ClpNode array is 2**N long. Values for N and array are in stuff (nNodes_ also in stuff)

Double checks OK.

Starts Fast dual2.

Like Fast dual.

Stops Fast dual2.

ClpSimplex* ClpSimplex::fastCrunch ( ClpNodeStuff stuff,
int  mode 
)

Deals with crunch aspects mode 0 - in 1 - out with solution 2 - out without solution returns small model or NULL.

Pivot in a variable and out a variable.

Returns 0 if okay, 1 if inaccuracy forced re-factorization, -1 if would be singular. Also updates primal/dual infeasibilities. Assumes sequenceIn_ and pivotRow_ set and also directionIn and Out.

Pivot in a variable and choose an outgoing one.

Assumes primal feasible - will not go through a bound. Returns step length in theta Returns ray in ray_ (or NULL if no pivot) Return codes as before but -1 means no acceptable pivot

Pivot out a variable and choose an incoing one.

Assumes dual feasible - will not go through a reduced cost. Returns step length in theta Returns ray in ray_ (or NULL if no pivot) Return codes as before but -1 means no acceptable pivot

int ClpSimplex::startup ( int  ifValuesPass,
int  startFinishOptions = 0 
)

Common bits of coding for dual and primal.

Return 0 if okay, 1 if bad matrix, 2 if very bad factorization

startFinishOptions - bits 1 - do not delete work areas and factorization at end 2 - use old factorization if same number of rows 4 - skip as much initialization of work areas as possible (based on whatsChanged in clpmodel.hpp) ** work in progress maybe other bits later

void ClpSimplex::finish ( int  startFinishOptions = 0)

Pivot in a variable and out a variable.

Returns 0 if okay, 1 if inaccuracy forced re-factorization, -1 if would be singular. Also updates primal/dual infeasibilities. Assumes sequenceIn_ and pivotRow_ set and also directionIn and Out.

bool ClpSimplex::statusOfProblem ( bool  initial = false)

Factorizes and returns true if optimal.

Used by user

If user left factorization frequency then compute.

bool ClpSimplex::primalFeasible ( ) const [inline]

If problem is primal feasible.

Definition at line 488 of file ClpSimplex.hpp.

bool ClpSimplex::dualFeasible ( ) const [inline]

If problem is dual feasible.

Definition at line 492 of file ClpSimplex.hpp.

factorization

Definition at line 496 of file ClpSimplex.hpp.

Sparsity on or off.

void ClpSimplex::setSparseFactorization ( bool  value)

If problem is primal feasible.

Factorization frequency.

If problem is primal feasible.

double ClpSimplex::dualBound ( ) const [inline]

Dual bound.

Definition at line 506 of file ClpSimplex.hpp.

void ClpSimplex::setDualBound ( double  value)

If problem is primal feasible.

double ClpSimplex::infeasibilityCost ( ) const [inline]

Infeasibility cost.

Definition at line 511 of file ClpSimplex.hpp.

void ClpSimplex::setInfeasibilityCost ( double  value)

If problem is primal feasible.

int ClpSimplex::perturbation ( ) const [inline]

Amount of print out: 0 - none 1 - just final 2 - just factorizations 3 - as 2 plus a bit more 4 - verbose above that 8,16,32 etc just for selective debug.

Perturbation: 50 - switch on perturbation 100 - auto perturb if takes too long (1.0e-6 largest nonzero) 101 - we are perturbed 102 - don't try perturbing again default is 100 others are for playing

Definition at line 531 of file ClpSimplex.hpp.

void ClpSimplex::setPerturbation ( int  value)

If problem is primal feasible.

int ClpSimplex::algorithm ( ) const [inline]

Current (or last) algorithm.

Definition at line 536 of file ClpSimplex.hpp.

void ClpSimplex::setAlgorithm ( int  value) [inline]

Set algorithm.

Definition at line 540 of file ClpSimplex.hpp.

Return true if the objective limit test can be relied upon.

double ClpSimplex::sumDualInfeasibilities ( ) const [inline]

Sum of dual infeasibilities.

Definition at line 546 of file ClpSimplex.hpp.

void ClpSimplex::setSumDualInfeasibilities ( double  value) [inline]

If problem is primal feasible.

Definition at line 549 of file ClpSimplex.hpp.

Sum of relaxed dual infeasibilities.

Definition at line 553 of file ClpSimplex.hpp.

void ClpSimplex::setSumOfRelaxedDualInfeasibilities ( double  value) [inline]

If problem is primal feasible.

Definition at line 556 of file ClpSimplex.hpp.

int ClpSimplex::numberDualInfeasibilities ( ) const [inline]

Number of dual infeasibilities.

Definition at line 560 of file ClpSimplex.hpp.

void ClpSimplex::setNumberDualInfeasibilities ( int  value) [inline]

If problem is primal feasible.

Definition at line 563 of file ClpSimplex.hpp.

Number of dual infeasibilities (without free)

Definition at line 567 of file ClpSimplex.hpp.

double ClpSimplex::sumPrimalInfeasibilities ( ) const [inline]

Sum of primal infeasibilities.

Definition at line 571 of file ClpSimplex.hpp.

void ClpSimplex::setSumPrimalInfeasibilities ( double  value) [inline]

If problem is primal feasible.

Definition at line 574 of file ClpSimplex.hpp.

Sum of relaxed primal infeasibilities.

Definition at line 578 of file ClpSimplex.hpp.

void ClpSimplex::setSumOfRelaxedPrimalInfeasibilities ( double  value) [inline]

If problem is primal feasible.

Definition at line 581 of file ClpSimplex.hpp.

Number of primal infeasibilities.

Definition at line 585 of file ClpSimplex.hpp.

void ClpSimplex::setNumberPrimalInfeasibilities ( int  value) [inline]

If problem is primal feasible.

Definition at line 588 of file ClpSimplex.hpp.

int ClpSimplex::saveModel ( const char *  fileName)

Save model to file, returns 0 if success.

This is designed for use outside algorithms so does not save iterating arrays etc. It does not save any messaging information. Does not save scaling values. It does not know about all types of virtual functions.

int ClpSimplex::restoreModel ( const char *  fileName)

Restore model from file, returns 0 if success, deletes current model.

void ClpSimplex::checkSolution ( int  setToBounds = 0)

Just check solution (for external use) - sets sum of infeasibilities etc.

If setToBounds 0 then primal column values not changed and used to compute primal row activity values. If 1 or 2 then status used - so all nonbasic variables set to indicated bound and if any values changed (or ==2) basic values re-computed.

Just check solution (for internal use) - sets sum of infeasibilities etc.

CoinIndexedVector* ClpSimplex::rowArray ( int  index) const [inline]

Useful row length arrays (0,1,2,3,4,5)

Definition at line 614 of file ClpSimplex.hpp.

CoinIndexedVector* ClpSimplex::columnArray ( int  index) const [inline]

Useful column length arrays (0,1,2,3,4,5)

Definition at line 618 of file ClpSimplex.hpp.

int ClpSimplex::getSolution ( const double *  rowActivities,
const double *  columnActivities 
)

Given an existing factorization computes and checks primal and dual solutions.

Uses input arrays for variables at bounds. Returns feasibility states

Given an existing factorization computes and checks primal and dual solutions.

Uses current problem arrays for bounds. Returns feasibility states

int ClpSimplex::createPiecewiseLinearCosts ( const int *  starts,
const double *  lower,
const double *  gradient 
)

Constructs a non linear cost from list of non-linearities (columns only) First lower of each column is taken as real lower Last lower is taken as real upper and cost ignored.

Returns nonzero if bad data e.g. lowers not monotonic

dual row pivot choice

Definition at line 644 of file ClpSimplex.hpp.

primal column pivot choice

Definition at line 648 of file ClpSimplex.hpp.

bool ClpSimplex::goodAccuracy ( ) const [inline]

Returns true if model looks OK.

Definition at line 652 of file ClpSimplex.hpp.

void ClpSimplex::returnModel ( ClpSimplex otherModel)

Return model - updates any scalars.

int ClpSimplex::internalFactorize ( int  solveType)

Factorizes using current basis.

solveType - 1 iterating, 0 initial, -1 external If 10 added then in primal values pass Return codes are as from ClpFactorization unless initial factorization when total number of singularities is returned. Special case is numberRows_+1 -> all slack basis.

Save data.

Restore data.

Clean up status.

Factorizes using current basis. For external use.

void ClpSimplex::computeDuals ( double *  givenDjs)

Computes duals from scratch.

If givenDjs then allows for nonzero basic djs

void ClpSimplex::computePrimals ( const double *  rowActivities,
const double *  columnActivities 
)

Computes primals from scratch.

void ClpSimplex::add ( double *  array,
int  column,
double  multiplier 
) const

Adds multiple of a column into an array.

void ClpSimplex::unpack ( CoinIndexedVector *  rowArray) const

Unpacks one column of the matrix into indexed array Uses sequenceIn_ Also applies scaling if needed.

void ClpSimplex::unpack ( CoinIndexedVector *  rowArray,
int  sequence 
) const

Unpacks one column of the matrix into indexed array Slack if sequence>= numberColumns Also applies scaling if needed.

void ClpSimplex::unpackPacked ( CoinIndexedVector *  rowArray)

Unpacks one column of the matrix into indexed array as packed vector Uses sequenceIn_ Also applies scaling if needed.

void ClpSimplex::unpackPacked ( CoinIndexedVector *  rowArray,
int  sequence 
)

Unpacks one column of the matrix into indexed array as packed vector Slack if sequence>= numberColumns Also applies scaling if needed.

int ClpSimplex::housekeeping ( double  objectiveChange) [protected]

This does basis housekeeping and does values for in/out variables.

Can also decide to re-factorize

void ClpSimplex::checkPrimalSolution ( const double *  rowActivities = NULL,
const double *  columnActivies = NULL 
) [protected]

This sets largest infeasibility and most infeasible and sum and number of infeasibilities (Primal)

void ClpSimplex::checkDualSolution ( ) [protected]

This sets largest infeasibility and most infeasible and sum and number of infeasibilities (Dual)

void ClpSimplex::checkBothSolutions ( ) [protected]

This sets sum and number of infeasibilities (Dual and Primal)

double ClpSimplex::scaleObjective ( double  value) [protected]

If input negative scales objective so maximum <= -value and returns scale factor used.

If positive unscales and also redoes dual stuff

int ClpSimplex::solveDW ( CoinStructuredModel *  model) [protected]

Solve using Dantzig-Wolfe decomposition and maybe in parallel.

int ClpSimplex::solveBenders ( CoinStructuredModel *  model) [protected]

Solve using Benders decomposition and maybe in parallel.

void ClpSimplex::setValuesPassAction ( double  incomingInfeasibility,
double  allowedInfeasibility 
)

For advanced use.

When doing iterative solves things can get nasty so on values pass if incoming solution has largest infeasibility < incomingInfeasibility throw out variables from basis until largest infeasibility < allowedInfeasibility or incoming largest infeasibility. If allowedInfeasibility>= incomingInfeasibility this is always possible altough you may end up with an all slack basis.

Defaults are 1.0,10.0

double ClpSimplex::alphaAccuracy ( ) const [inline]

Initial value for alpha accuracy calculation (-1.0 off)

Definition at line 750 of file ClpSimplex.hpp.

void ClpSimplex::setAlphaAccuracy ( double  value) [inline]

If problem is primal feasible.

Definition at line 753 of file ClpSimplex.hpp.

void ClpSimplex::setDisasterHandler ( ClpDisasterHandler handler) [inline]

Objective value.

Set disaster handler

Definition at line 762 of file ClpSimplex.hpp.

Get disaster handler.

Definition at line 766 of file ClpSimplex.hpp.

double ClpSimplex::largeValue ( ) const [inline]

Large bound value (for complementarity etc)

Definition at line 770 of file ClpSimplex.hpp.

void ClpSimplex::setLargeValue ( double  value)

If problem is primal feasible.

double ClpSimplex::largestPrimalError ( ) const [inline]

Largest error on Ax-b.

Definition at line 775 of file ClpSimplex.hpp.

double ClpSimplex::largestDualError ( ) const [inline]

Largest error on basic duals.

Definition at line 779 of file ClpSimplex.hpp.

void ClpSimplex::setLargestPrimalError ( double  value) [inline]

Largest error on Ax-b.

Definition at line 783 of file ClpSimplex.hpp.

void ClpSimplex::setLargestDualError ( double  value) [inline]

Largest error on basic duals.

Definition at line 787 of file ClpSimplex.hpp.

double ClpSimplex::zeroTolerance ( ) const [inline]

Get zero tolerance.

Definition at line 791 of file ClpSimplex.hpp.

void ClpSimplex::setZeroTolerance ( double  value) [inline]

Set zero tolerance.

Definition at line 795 of file ClpSimplex.hpp.

int* ClpSimplex::pivotVariable ( ) const [inline]

Basic variables pivoting on which rows.

Definition at line 799 of file ClpSimplex.hpp.

bool ClpSimplex::automaticScaling ( ) const [inline]

If automatic scaling on.

Definition at line 803 of file ClpSimplex.hpp.

void ClpSimplex::setAutomaticScaling ( bool  onOff) [inline]

If problem is primal feasible.

Definition at line 806 of file ClpSimplex.hpp.

double ClpSimplex::currentDualTolerance ( ) const [inline]

Current dual tolerance.

Definition at line 810 of file ClpSimplex.hpp.

void ClpSimplex::setCurrentDualTolerance ( double  value) [inline]

If problem is primal feasible.

Definition at line 813 of file ClpSimplex.hpp.

double ClpSimplex::currentPrimalTolerance ( ) const [inline]

Current primal tolerance.

Definition at line 817 of file ClpSimplex.hpp.

void ClpSimplex::setCurrentPrimalTolerance ( double  value) [inline]

If problem is primal feasible.

Definition at line 820 of file ClpSimplex.hpp.

int ClpSimplex::numberRefinements ( ) const [inline]

How many iterative refinements to do.

Definition at line 824 of file ClpSimplex.hpp.

void ClpSimplex::setNumberRefinements ( int  value)

If problem is primal feasible.

double ClpSimplex::alpha ( ) const [inline]

Alpha (pivot element) for use by classes e.g. steepestedge.

Definition at line 829 of file ClpSimplex.hpp.

void ClpSimplex::setAlpha ( double  value) [inline]

If problem is primal feasible.

Definition at line 832 of file ClpSimplex.hpp.

double ClpSimplex::dualIn ( ) const [inline]

Reduced cost of last incoming for use by classes e.g. steepestedge.

Definition at line 836 of file ClpSimplex.hpp.

int ClpSimplex::pivotRow ( ) const [inline]

Pivot Row for use by classes e.g. steepestedge.

Definition at line 840 of file ClpSimplex.hpp.

void ClpSimplex::setPivotRow ( int  value) [inline]

If problem is primal feasible.

Definition at line 843 of file ClpSimplex.hpp.

value of incoming variable (in Dual)

int ClpSimplex::gutsOfSolution ( double *  givenDuals,
const double *  givenPrimals,
bool  valuesPass = false 
) [protected]

May change basis and then returns number changed.

Computation of solutions may be overriden by given pi and solution

void ClpSimplex::gutsOfDelete ( int  type) [protected]

Does most of deletion (0 = all, 1 = most, 2 most + factorization)

Reimplemented from ClpModel.

void ClpSimplex::gutsOfCopy ( const ClpSimplex rhs) [protected]

Does most of copying.

bool ClpSimplex::createRim ( int  what,
bool  makeRowCopy = false,
int  startFinishOptions = 0 
) [protected]

puts in format I like (rowLower,rowUpper) also see StandardMatrix 1 bit does rows (now and columns), (2 bit does column bounds), 4 bit does objective(s).

8 bit does solution scaling in 16 bit does rowArray and columnArray indexed vectors and makes row copy if wanted, also sets columnStart_ etc Also creates scaling arrays if needed. It does scaling if needed. 16 also moves solutions etc in to work arrays On 16 returns false if problem "bad" i.e. matrix or bounds bad If startFinishOptions is -1 then called by user in getSolution so do arrays but keep pivotVariable_

void ClpSimplex::createRim1 ( bool  initial) [protected]

Does rows and columns.

void ClpSimplex::createRim4 ( bool  initial) [protected]

Does objective.

void ClpSimplex::createRim5 ( bool  initial) [protected]

Does rows and columns and objective.

void ClpSimplex::deleteRim ( int  getRidOfFactorizationData = 2) [protected]

releases above arrays and does solution scaling out.

May also get rid of factorization data - 0 get rid of nothing, 1 get rid of arrays, 2 also factorization

bool ClpSimplex::sanityCheck ( ) [protected]

Sanity check on input rim data (after scaling) - returns true if okay.

double* ClpSimplex::solutionRegion ( int  section) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 894 of file ClpSimplex.hpp.

double* ClpSimplex::djRegion ( int  section) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 898 of file ClpSimplex.hpp.

double* ClpSimplex::lowerRegion ( int  section) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 902 of file ClpSimplex.hpp.

double* ClpSimplex::upperRegion ( int  section) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 906 of file ClpSimplex.hpp.

double* ClpSimplex::costRegion ( int  section) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 910 of file ClpSimplex.hpp.

double* ClpSimplex::solutionRegion ( ) const [inline]

Return region as single array.

Definition at line 915 of file ClpSimplex.hpp.

double* ClpSimplex::djRegion ( ) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 918 of file ClpSimplex.hpp.

double* ClpSimplex::lowerRegion ( ) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 921 of file ClpSimplex.hpp.

double* ClpSimplex::upperRegion ( ) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 924 of file ClpSimplex.hpp.

double* ClpSimplex::costRegion ( ) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 927 of file ClpSimplex.hpp.

Status ClpSimplex::getStatus ( int  sequence) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 930 of file ClpSimplex.hpp.

void ClpSimplex::setStatus ( int  sequence,
Status  newstatus 
) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 933 of file ClpSimplex.hpp.

Start or reset using maximumRows_ and Columns_ - true if change.

Reimplemented from ClpModel.

Normally the first factorization does sparse coding because the factorization could be singular.

This allows initial dense factorization when it is known to be safe

Return row or column sections - not as much needed as it once was.

These just map into single arrays

int ClpSimplex::sequenceIn ( ) const [inline]

Return sequence In or Out.

Definition at line 947 of file ClpSimplex.hpp.

int ClpSimplex::sequenceOut ( ) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 950 of file ClpSimplex.hpp.

void ClpSimplex::setSequenceIn ( int  sequence) [inline]

Set sequenceIn or Out.

Definition at line 954 of file ClpSimplex.hpp.

void ClpSimplex::setSequenceOut ( int  sequence) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 957 of file ClpSimplex.hpp.

int ClpSimplex::directionIn ( ) const [inline]

Return direction In or Out.

Definition at line 961 of file ClpSimplex.hpp.

int ClpSimplex::directionOut ( ) const [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 964 of file ClpSimplex.hpp.

void ClpSimplex::setDirectionIn ( int  direction) [inline]

Set directionIn or Out.

Definition at line 968 of file ClpSimplex.hpp.

void ClpSimplex::setDirectionOut ( int  direction) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 971 of file ClpSimplex.hpp.

double ClpSimplex::valueOut ( ) const [inline]

Value of Out variable.

Definition at line 975 of file ClpSimplex.hpp.

void ClpSimplex::setValueOut ( double  value) [inline]

Set value of out variable.

Definition at line 979 of file ClpSimplex.hpp.

void ClpSimplex::setLowerOut ( double  value) [inline]

Set lower of out variable.

Definition at line 983 of file ClpSimplex.hpp.

void ClpSimplex::setUpperOut ( double  value) [inline]

Set upper of out variable.

Definition at line 987 of file ClpSimplex.hpp.

void ClpSimplex::setTheta ( double  value) [inline]

Set theta of out variable.

Definition at line 991 of file ClpSimplex.hpp.

int ClpSimplex::isColumn ( int  sequence) const [inline]

Returns 1 if sequence indicates column.

Definition at line 995 of file ClpSimplex.hpp.

int ClpSimplex::sequenceWithin ( int  sequence) const [inline]

Returns sequence number within section.

Definition at line 999 of file ClpSimplex.hpp.

double ClpSimplex::solution ( int  sequence) [inline]

Return row or column values.

Definition at line 1003 of file ClpSimplex.hpp.

double& ClpSimplex::solutionAddress ( int  sequence) [inline]

Return address of row or column values.

Definition at line 1007 of file ClpSimplex.hpp.

double ClpSimplex::reducedCost ( int  sequence) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 1010 of file ClpSimplex.hpp.

double& ClpSimplex::reducedCostAddress ( int  sequence) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 1013 of file ClpSimplex.hpp.

double ClpSimplex::lower ( int  sequence) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 1016 of file ClpSimplex.hpp.

double& ClpSimplex::lowerAddress ( int  sequence) [inline]

Return address of row or column lower bound.

Definition at line 1020 of file ClpSimplex.hpp.

double ClpSimplex::upper ( int  sequence) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 1023 of file ClpSimplex.hpp.

double& ClpSimplex::upperAddress ( int  sequence) [inline]

Return address of row or column upper bound.

Definition at line 1027 of file ClpSimplex.hpp.

double ClpSimplex::cost ( int  sequence) [inline]

Return row or column sections - not as much needed as it once was.

These just map into single arrays

Definition at line 1030 of file ClpSimplex.hpp.

double& ClpSimplex::costAddress ( int  sequence) [inline]

Return address of row or column cost.

Definition at line 1034 of file ClpSimplex.hpp.

double ClpSimplex::originalLower ( int  iSequence) const [inline]

Return original lower bound.

Definition at line 1038 of file ClpSimplex.hpp.

double ClpSimplex::originalUpper ( int  iSequence) const [inline]

Return original lower bound.

Definition at line 1044 of file ClpSimplex.hpp.

double ClpSimplex::theta ( ) const [inline]

Theta (pivot change)

Definition at line 1050 of file ClpSimplex.hpp.

double ClpSimplex::bestPossibleImprovement ( ) const [inline]

Best possible improvement using djs (primal) or obj change by flipping bounds to make dual feasible (dual)

Definition at line 1055 of file ClpSimplex.hpp.

Return pointer to details of costs.

Definition at line 1059 of file ClpSimplex.hpp.

int ClpSimplex::moreSpecialOptions ( ) const [inline]

Return more special options 1 bit - if presolve says infeasible in ClpSolve return 2 bit - if presolved problem infeasible return 4 bit - keep arrays like upper_ around 8 bit - if factorization kept can still declare optimal at once 16 bit - if checking replaceColumn accuracy before updating 32 bit - say optimal if primal feasible! 64 bit - give up easily in dual (and say infeasible) 128 bit - no objective, 0-1 and in B&B 256 bit - in primal from dual or vice versa 512 bit - alternative use of solveType_.

Definition at line 1074 of file ClpSimplex.hpp.

void ClpSimplex::setMoreSpecialOptions ( int  value) [inline]

Set more special options 1 bit - if presolve says infeasible in ClpSolve return 2 bit - if presolved problem infeasible return 4 bit - keep arrays like upper_ around 8 bit - no free or superBasic variables 16 bit - if checking replaceColumn accuracy before updating 32 bit - say optimal if primal feasible! 64 bit - give up easily in dual (and say infeasible) 128 bit - no objective, 0-1 and in B&B 256 bit - in primal from dual or vice versa 512 bit - alternative use of solveType_.

Definition at line 1089 of file ClpSimplex.hpp.

void ClpSimplex::setFakeBound ( int  sequence,
FakeBound  fakeBound 
) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1095 of file ClpSimplex.hpp.

FakeBound ClpSimplex::getFakeBound ( int  sequence) const [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1100 of file ClpSimplex.hpp.

void ClpSimplex::setRowStatus ( int  sequence,
Status  newstatus 
) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1103 of file ClpSimplex.hpp.

Status ClpSimplex::getRowStatus ( int  sequence) const [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1108 of file ClpSimplex.hpp.

void ClpSimplex::setColumnStatus ( int  sequence,
Status  newstatus 
) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1111 of file ClpSimplex.hpp.

Status ClpSimplex::getColumnStatus ( int  sequence) const [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1116 of file ClpSimplex.hpp.

void ClpSimplex::setPivoted ( int  sequence) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1119 of file ClpSimplex.hpp.

void ClpSimplex::clearPivoted ( int  sequence) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1122 of file ClpSimplex.hpp.

bool ClpSimplex::pivoted ( int  sequence) const [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1125 of file ClpSimplex.hpp.

void ClpSimplex::setFlagged ( int  sequence)

To flag a variable (not inline to allow for column generation)

void ClpSimplex::clearFlagged ( int  sequence) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1130 of file ClpSimplex.hpp.

bool ClpSimplex::flagged ( int  sequence) const [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1133 of file ClpSimplex.hpp.

void ClpSimplex::setActive ( int  iRow) [inline]

To say row active in primal pivot row choice.

Definition at line 1137 of file ClpSimplex.hpp.

void ClpSimplex::clearActive ( int  iRow) [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1140 of file ClpSimplex.hpp.

bool ClpSimplex::active ( int  iRow) const [inline]

To flag a variable (not inline to allow for column generation)

Definition at line 1143 of file ClpSimplex.hpp.

Set up status array (can be used by OsiClp).

Also can be used to set up all slack basis

void ClpSimplex::allSlackBasis ( bool  resetSolution = false)

Sets up all slack basis and resets solution to as it was after initial load or readMps.

int ClpSimplex::lastBadIteration ( ) const [inline]

So we know when to be cautious.

Definition at line 1154 of file ClpSimplex.hpp.

int ClpSimplex::progressFlag ( ) const [inline]

Progress flag - at present 0 bit says artificials out.

Definition at line 1158 of file ClpSimplex.hpp.

void ClpSimplex::forceFactorization ( int  value) [inline]

Force re-factorization early.

Definition at line 1162 of file ClpSimplex.hpp.

double ClpSimplex::rawObjectiveValue ( ) const [inline]

Raw objective value (so always minimize in primal)

Reimplemented from ClpModel.

Definition at line 1166 of file ClpSimplex.hpp.

void ClpSimplex::computeObjectiveValue ( bool  useWorkingSolution = false)

Compute objective value from solution and put in objectiveValue_.

Compute minimization objective value from internal solution without perturbation.

int ClpSimplex::numberExtraRows ( ) const [inline]

Number of extra rows.

These are ones which will be dynamically created each iteration. This is for GUB but may have other uses.

Definition at line 1176 of file ClpSimplex.hpp.

int ClpSimplex::maximumBasic ( ) const [inline]

Maximum number of basic variables - can be more than number of rows if GUB.

Definition at line 1181 of file ClpSimplex.hpp.

int ClpSimplex::baseIteration ( ) const [inline]

Iteration when we entered dual or primal.

Definition at line 1185 of file ClpSimplex.hpp.

void ClpSimplex::generateCpp ( FILE *  fp,
bool  defaultFactor = false 
)

Create C++ lines to get to current state.

Gets clean and emptyish factorization.

May delete or may make clean and emptyish factorization.

void ClpSimplex::moveInfo ( const ClpSimplex rhs,
bool  justStatus = false 
)

Move status and solution across.

void ClpSimplex::getBInvARow ( int  row,
double *  z,
double *  slack = NULL 
)

Get a row of the tableau (slack part in slack if not NULL)

void ClpSimplex::getBInvRow ( int  row,
double *  z 
)

Get a row of the basis inverse.

void ClpSimplex::getBInvACol ( int  col,
double *  vec 
)

Get a column of the tableau.

void ClpSimplex::getBInvCol ( int  col,
double *  vec 
)

Get a column of the basis inverse.

void ClpSimplex::getBasics ( int *  index)

Get basic indices (order of indices corresponds to the order of elements in a vector retured by getBInvACol() and getBInvCol()).

void ClpSimplex::setObjectiveCoefficient ( int  elementIndex,
double  elementValue 
)

Set an objective function coefficient.

Reimplemented from ClpModel.

void ClpSimplex::setObjCoeff ( int  elementIndex,
double  elementValue 
) [inline]

Set an objective function coefficient.

Reimplemented from ClpModel.

Definition at line 1227 of file ClpSimplex.hpp.

void ClpSimplex::setColumnLower ( int  elementIndex,
double  elementValue 
)

Set a single column lower bound
Use -DBL_MAX for -infinity.

Reimplemented from ClpModel.

void ClpSimplex::setColumnUpper ( int  elementIndex,
double  elementValue 
)

Set a single column upper bound
Use DBL_MAX for infinity.

Reimplemented from ClpModel.

void ClpSimplex::setColumnBounds ( int  elementIndex,
double  lower,
double  upper 
)

Set a single column lower and upper bound.

Reimplemented from ClpModel.

void ClpSimplex::setColumnSetBounds ( const int *  indexFirst,
const int *  indexLast,
const double *  boundList 
)

Set the bounds on a number of columns simultaneously
The default implementation just invokes setColLower() and setColUpper() over and over again.

Parameters:
indexFirst,indexLastpointers to the beginning and after the end of the array of the indices of the variables whose either bound changes
boundListthe new lower/upper bound pairs for the variables

Reimplemented from ClpModel.

void ClpSimplex::setColLower ( int  elementIndex,
double  elementValue 
) [inline]

Set a single column lower bound
Use -DBL_MAX for -infinity.

Reimplemented from ClpModel.

Definition at line 1257 of file ClpSimplex.hpp.

void ClpSimplex::setColUpper ( int  elementIndex,
double  elementValue 
) [inline]

Set a single column upper bound
Use DBL_MAX for infinity.

Reimplemented from ClpModel.

Definition at line 1262 of file ClpSimplex.hpp.

void ClpSimplex::setColBounds ( int  elementIndex,
double  newlower,
double  newupper 
) [inline]

Set a single column lower and upper bound.

Reimplemented from ClpModel.

Definition at line 1267 of file ClpSimplex.hpp.

void ClpSimplex::setColSetBounds ( const int *  indexFirst,
const int *  indexLast,
const double *  boundList 
) [inline]

Set the bounds on a number of columns simultaneously

Parameters:
indexFirst,indexLastpointers to the beginning and after the end of the array of the indices of the variables whose either bound changes
boundListthe new lower/upper bound pairs for the variables

Reimplemented from ClpModel.

Definition at line 1278 of file ClpSimplex.hpp.

void ClpSimplex::setRowLower ( int  elementIndex,
double  elementValue 
)

Set a single row lower bound
Use -DBL_MAX for -infinity.

Reimplemented from ClpModel.

void ClpSimplex::setRowUpper ( int  elementIndex,
double  elementValue 
)

Set a single row upper bound
Use DBL_MAX for infinity.

Reimplemented from ClpModel.

void ClpSimplex::setRowBounds ( int  elementIndex,
double  lower,
double  upper 
)

Set a single row lower and upper bound.

Reimplemented from ClpModel.

void ClpSimplex::setRowSetBounds ( const int *  indexFirst,
const int *  indexLast,
const double *  boundList 
)

Set the bounds on a number of rows simultaneously

Parameters:
indexFirst,indexLastpointers to the beginning and after the end of the array of the indices of the constraints whose either bound changes
boundListthe new lower/upper bound pairs for the constraints

Reimplemented from ClpModel.

void ClpSimplex::resize ( int  newNumberRows,
int  newNumberColumns 
)

Resizes rim part of model.

Reimplemented from ClpModel.


Friends And Related Function Documentation

friend class OsiClpSolverInterface [friend]

Allow OsiClp certain perks.

Definition at line 1521 of file ClpSimplex.hpp.

void ClpSimplexUnitTest ( const std::string &  mpsDir) [friend]

A function that tests the methods in the ClpSimplex class.

The only reason for it not to be a member method is that this way it doesn't have to be compiled into the library. And that's a gain, because the library should be compiled with optimization on, but this method should be compiled with debugging.

It also does some testing of ClpFactorization class


Member Data Documentation

Best possible improvement using djs (primal) or obj change by flipping bounds to make dual feasible (dual)

Definition at line 1321 of file ClpSimplex.hpp.

double ClpSimplex::zeroTolerance_ [protected]

Zero tolerance.

Definition at line 1323 of file ClpSimplex.hpp.

Sequence of worst (-1 if feasible)

Definition at line 1325 of file ClpSimplex.hpp.

Sequence of worst (-1 if feasible)

Definition at line 1327 of file ClpSimplex.hpp.

double ClpSimplex::bestObjectiveValue_ [protected]

"Best" objective value

Definition at line 1329 of file ClpSimplex.hpp.

More special options - see set for details.

Definition at line 1331 of file ClpSimplex.hpp.

int ClpSimplex::baseIteration_ [protected]

Iteration when we entered dual or primal.

Definition at line 1333 of file ClpSimplex.hpp.

Primal tolerance needed to make dual feasible (<largeTolerance)

Definition at line 1335 of file ClpSimplex.hpp.

double ClpSimplex::largeValue_ [protected]

Large bound value (for complementarity etc)

Definition at line 1337 of file ClpSimplex.hpp.

double ClpSimplex::largestPrimalError_ [protected]

Largest error on Ax-b.

Definition at line 1339 of file ClpSimplex.hpp.

double ClpSimplex::largestDualError_ [protected]

Largest error on basic duals.

Definition at line 1341 of file ClpSimplex.hpp.

double ClpSimplex::alphaAccuracy_ [protected]

For computing whether to re-factorize.

Definition at line 1343 of file ClpSimplex.hpp.

double ClpSimplex::dualBound_ [protected]

Dual bound.

Definition at line 1345 of file ClpSimplex.hpp.

double ClpSimplex::alpha_ [protected]

Alpha (pivot element)

Definition at line 1347 of file ClpSimplex.hpp.

double ClpSimplex::theta_ [protected]

Theta (pivot change)

Definition at line 1349 of file ClpSimplex.hpp.

double ClpSimplex::lowerIn_ [protected]

Lower Bound on In variable.

Definition at line 1351 of file ClpSimplex.hpp.

double ClpSimplex::valueIn_ [protected]

Value of In variable.

Definition at line 1353 of file ClpSimplex.hpp.

double ClpSimplex::upperIn_ [protected]

Upper Bound on In variable.

Definition at line 1355 of file ClpSimplex.hpp.

double ClpSimplex::dualIn_ [protected]

Reduced cost of In variable.

Definition at line 1357 of file ClpSimplex.hpp.

double ClpSimplex::lowerOut_ [protected]

Lower Bound on Out variable.

Definition at line 1359 of file ClpSimplex.hpp.

double ClpSimplex::valueOut_ [protected]

Value of Out variable.

Definition at line 1361 of file ClpSimplex.hpp.

double ClpSimplex::upperOut_ [protected]

Upper Bound on Out variable.

Definition at line 1363 of file ClpSimplex.hpp.

double ClpSimplex::dualOut_ [protected]

Infeasibility (dual) or ? (primal) of Out variable.

Definition at line 1365 of file ClpSimplex.hpp.

double ClpSimplex::dualTolerance_ [protected]

Current dual tolerance for algorithm.

Definition at line 1367 of file ClpSimplex.hpp.

double ClpSimplex::primalTolerance_ [protected]

Current primal tolerance for algorithm.

Definition at line 1369 of file ClpSimplex.hpp.

Sum of dual infeasibilities.

Definition at line 1371 of file ClpSimplex.hpp.

Sum of primal infeasibilities.

Definition at line 1373 of file ClpSimplex.hpp.

double ClpSimplex::infeasibilityCost_ [protected]

Weight assigned to being infeasible in primal.

Definition at line 1375 of file ClpSimplex.hpp.

Sum of Dual infeasibilities using tolerance based on error in duals.

Definition at line 1377 of file ClpSimplex.hpp.

Sum of Primal infeasibilities using tolerance based on error in primals.

Definition at line 1379 of file ClpSimplex.hpp.

double ClpSimplex::acceptablePivot_ [protected]

Acceptable pivot value just after factorization.

Definition at line 1381 of file ClpSimplex.hpp.

double* ClpSimplex::lower_ [protected]

Working copy of lower bounds (Owner of arrays below)

Definition at line 1383 of file ClpSimplex.hpp.

double* ClpSimplex::rowLowerWork_ [protected]

Row lower bounds - working copy.

Definition at line 1385 of file ClpSimplex.hpp.

double* ClpSimplex::columnLowerWork_ [protected]

Column lower bounds - working copy.

Definition at line 1387 of file ClpSimplex.hpp.

double* ClpSimplex::upper_ [protected]

Working copy of upper bounds (Owner of arrays below)

Definition at line 1389 of file ClpSimplex.hpp.

double* ClpSimplex::rowUpperWork_ [protected]

Row upper bounds - working copy.

Definition at line 1391 of file ClpSimplex.hpp.

double* ClpSimplex::columnUpperWork_ [protected]

Column upper bounds - working copy.

Definition at line 1393 of file ClpSimplex.hpp.

double* ClpSimplex::cost_ [protected]

Working copy of objective (Owner of arrays below)

Definition at line 1395 of file ClpSimplex.hpp.

double* ClpSimplex::rowObjectiveWork_ [protected]

Row objective - working copy.

Definition at line 1397 of file ClpSimplex.hpp.

double* ClpSimplex::objectiveWork_ [protected]

Column objective - working copy.

Definition at line 1399 of file ClpSimplex.hpp.

CoinIndexedVector* ClpSimplex::rowArray_[6] [protected]

Useful row length arrays.

Definition at line 1401 of file ClpSimplex.hpp.

CoinIndexedVector* ClpSimplex::columnArray_[6] [protected]

Useful column length arrays.

Definition at line 1403 of file ClpSimplex.hpp.

int ClpSimplex::sequenceIn_ [protected]

Sequence of In variable.

Definition at line 1405 of file ClpSimplex.hpp.

int ClpSimplex::directionIn_ [protected]

Direction of In, 1 going up, -1 going down, 0 not a clude.

Definition at line 1407 of file ClpSimplex.hpp.

int ClpSimplex::sequenceOut_ [protected]

Sequence of Out variable.

Definition at line 1409 of file ClpSimplex.hpp.

int ClpSimplex::directionOut_ [protected]

Direction of Out, 1 to upper bound, -1 to lower bound, 0 - superbasic.

Definition at line 1411 of file ClpSimplex.hpp.

int ClpSimplex::pivotRow_ [protected]

Pivot Row.

Definition at line 1413 of file ClpSimplex.hpp.

Last good iteration (immediately after a re-factorization)

Definition at line 1415 of file ClpSimplex.hpp.

double* ClpSimplex::dj_ [protected]

Working copy of reduced costs (Owner of arrays below)

Definition at line 1417 of file ClpSimplex.hpp.

double* ClpSimplex::rowReducedCost_ [protected]

Reduced costs of slacks not same as duals (or - duals)

Definition at line 1419 of file ClpSimplex.hpp.

double* ClpSimplex::reducedCostWork_ [protected]

Possible scaled reduced costs.

Definition at line 1421 of file ClpSimplex.hpp.

double* ClpSimplex::solution_ [protected]

Working copy of primal solution (Owner of arrays below)

Definition at line 1423 of file ClpSimplex.hpp.

double* ClpSimplex::rowActivityWork_ [protected]

Row activities - working copy.

Definition at line 1425 of file ClpSimplex.hpp.

double* ClpSimplex::columnActivityWork_ [protected]

Column activities - working copy.

Definition at line 1427 of file ClpSimplex.hpp.

Number of dual infeasibilities.

Definition at line 1429 of file ClpSimplex.hpp.

Number of dual infeasibilities (without free)

Definition at line 1431 of file ClpSimplex.hpp.

Number of primal infeasibilities.

Definition at line 1433 of file ClpSimplex.hpp.

How many iterative refinements to do.

Definition at line 1435 of file ClpSimplex.hpp.

dual row pivot choice

Definition at line 1437 of file ClpSimplex.hpp.

primal column pivot choice

Definition at line 1439 of file ClpSimplex.hpp.

int* ClpSimplex::pivotVariable_ [protected]

Basic variables pivoting on which rows.

Definition at line 1441 of file ClpSimplex.hpp.

factorization

Definition at line 1443 of file ClpSimplex.hpp.

double* ClpSimplex::savedSolution_ [protected]

Saved version of solution.

Definition at line 1445 of file ClpSimplex.hpp.

Number of times code has tentatively thought optimal.

Definition at line 1447 of file ClpSimplex.hpp.

Disaster handler.

Definition at line 1449 of file ClpSimplex.hpp.

int ClpSimplex::changeMade_ [protected]

If change has been made (first attempt at stopping looping)

Definition at line 1451 of file ClpSimplex.hpp.

int ClpSimplex::algorithm_ [protected]

Algorithm >0 == Primal, <0 == Dual.

Definition at line 1453 of file ClpSimplex.hpp.

Now for some reliability aids This forces re-factorization early.

Definition at line 1456 of file ClpSimplex.hpp.

int ClpSimplex::perturbation_ [protected]

Perturbation:

-50 to +50 - perturb by this power of ten (-6 sounds good) 100 - auto perturb if takes too long (1.0e-6 largest nonzero) 101 - we are perturbed 102 - don't try perturbing again default is 100

Definition at line 1464 of file ClpSimplex.hpp.

unsigned char* ClpSimplex::saveStatus_ [protected]

Saved status regions.

Definition at line 1466 of file ClpSimplex.hpp.

Very wasteful way of dealing with infeasibilities in primal.

However it will allow non-linearities and use of dual analysis. If it doesn't work it can easily be replaced.

Definition at line 1471 of file ClpSimplex.hpp.

So we know when to be cautious.

Definition at line 1473 of file ClpSimplex.hpp.

So we know when to open up again.

Definition at line 1475 of file ClpSimplex.hpp.

int ClpSimplex::numberFake_ [protected]

Can be used for count of fake bounds (dual) or fake costs (primal)

Definition at line 1477 of file ClpSimplex.hpp.

int ClpSimplex::numberChanged_ [protected]

Can be used for count of changed costs (dual) or changed bounds (primal)

Definition at line 1479 of file ClpSimplex.hpp.

int ClpSimplex::progressFlag_ [protected]

Progress flag - at present 0 bit says artificials out, 1 free in.

Definition at line 1481 of file ClpSimplex.hpp.

int ClpSimplex::firstFree_ [protected]

First free/super-basic variable (-1 if none)

Definition at line 1483 of file ClpSimplex.hpp.

Number of extra rows.

These are ones which will be dynamically created each iteration. This is for GUB but may have other uses.

Definition at line 1487 of file ClpSimplex.hpp.

int ClpSimplex::maximumBasic_ [protected]

Maximum number of basic variables - can be more than number of rows if GUB.

Definition at line 1490 of file ClpSimplex.hpp.

If may skip final factorize then allow up to this pivots (default 20)

Definition at line 1492 of file ClpSimplex.hpp.

For advanced use.

When doing iterative solves things can get nasty so on values pass if incoming solution has largest infeasibility < incomingInfeasibility throw out variables from basis until largest infeasibility < allowedInfeasibility. if allowedInfeasibility>= incomingInfeasibility this is always possible altough you may end up with an all slack basis.

Defaults are 1.0,10.0

Definition at line 1502 of file ClpSimplex.hpp.

Best possible improvement using djs (primal) or obj change by flipping bounds to make dual feasible (dual)

Definition at line 1503 of file ClpSimplex.hpp.

int ClpSimplex::automaticScale_ [protected]

Automatic scaling of objective and rhs and bounds.

Definition at line 1505 of file ClpSimplex.hpp.

Maximum perturbation array size (take out when code rewritten)

Definition at line 1507 of file ClpSimplex.hpp.

double* ClpSimplex::perturbationArray_ [protected]

Perturbation array (maximumPerturbationSize_)

Definition at line 1509 of file ClpSimplex.hpp.

A copy of model with certain state - normally without cuts.

Definition at line 1511 of file ClpSimplex.hpp.

For dealing with all issues of cycling etc.

Definition at line 1513 of file ClpSimplex.hpp.

int ClpSimplex::spareIntArray_[4] [mutable]

Spare int array for passing information [0]!=0 switches on.

Definition at line 1516 of file ClpSimplex.hpp.

double ClpSimplex::spareDoubleArray_[4] [mutable]

Spare double array for passing information [0]!=0 switches on.

Definition at line 1518 of file ClpSimplex.hpp.


The documentation for this class was generated from the following file: