/* $Id: ClpPresolve.hpp 2134 2015-03-22 16:40:43Z forrest $ */ // Copyright (C) 2002, International Business Machines // Corporation and others. All Rights Reserved. // This code is licensed under the terms of the Eclipse Public License (EPL). #ifndef ClpPresolve_H #define ClpPresolve_H #include "ClpSimplex.hpp" class CoinPresolveAction; #include "CoinPresolveMatrix.hpp" /** This is the Clp interface to CoinPresolve */ class ClpPresolve { public: /**@name Main Constructor, destructor */ //@{ /// Default constructor ClpPresolve(); /// Virtual destructor virtual ~ClpPresolve(); //@} /**@name presolve - presolves a model, transforming the model * and saving information in the ClpPresolve object needed for postsolving. * This underlying (protected) method is virtual; the idea is that in the future, * one could override this method to customize how the various * presolve techniques are applied. This version of presolve returns a pointer to a new presolved model. NULL if infeasible or unbounded. This should be paired with postsolve below. The advantage of going back to original model is that it will be exactly as it was i.e. 0.0 will not become 1.0e-19. If keepIntegers is true then bounds may be tightened in original. Bounds will be moved by up to feasibilityTolerance to try and stay feasible. Names will be dropped in presolved model if asked */ ClpSimplex * presolvedModel(ClpSimplex & si, double feasibilityTolerance = 0.0, bool keepIntegers = true, int numberPasses = 5, bool dropNames = false, bool doRowObjective = false, const char * prohibitedRows=NULL, const char * prohibitedColumns=NULL); #ifndef CLP_NO_STD /** This version saves data in a file. The passed in model is updated to be presolved model. Returns non-zero if infeasible*/ int presolvedModelToFile(ClpSimplex &si, std::string fileName, double feasibilityTolerance = 0.0, bool keepIntegers = true, int numberPasses = 5, bool dropNames = false, bool doRowObjective = false); #endif /** Return pointer to presolved model, Up to user to destroy */ ClpSimplex * model() const; /// Return pointer to original model ClpSimplex * originalModel() const; /// Set pointer to original model void setOriginalModel(ClpSimplex * model); /// return pointer to original columns const int * originalColumns() const; /// return pointer to original rows const int * originalRows() const; /** "Magic" number. If this is non-zero then any elements with this value may change and so presolve is very limited in what can be done to the row and column. This is for non-linear problems. */ inline void setNonLinearValue(double value) { nonLinearValue_ = value; } inline double nonLinearValue() const { return nonLinearValue_; } /// Whether we want to do dual part of presolve inline bool doDual() const { return (presolveActions_ & 1) == 0; } inline void setDoDual(bool doDual) { if (doDual) presolveActions_ &= ~1; else presolveActions_ |= 1; } /// Whether we want to do singleton part of presolve inline bool doSingleton() const { return (presolveActions_ & 2) == 0; } inline void setDoSingleton(bool doSingleton) { if (doSingleton) presolveActions_ &= ~2; else presolveActions_ |= 2; } /// Whether we want to do doubleton part of presolve inline bool doDoubleton() const { return (presolveActions_ & 4) == 0; } inline void setDoDoubleton(bool doDoubleton) { if (doDoubleton) presolveActions_ &= ~4; else presolveActions_ |= 4; } /// Whether we want to do tripleton part of presolve inline bool doTripleton() const { return (presolveActions_ & 8) == 0; } inline void setDoTripleton(bool doTripleton) { if (doTripleton) presolveActions_ &= ~8; else presolveActions_ |= 8; } /// Whether we want to do tighten part of presolve inline bool doTighten() const { return (presolveActions_ & 16) == 0; } inline void setDoTighten(bool doTighten) { if (doTighten) presolveActions_ &= ~16; else presolveActions_ |= 16; } /// Whether we want to do forcing part of presolve inline bool doForcing() const { return (presolveActions_ & 32) == 0; } inline void setDoForcing(bool doForcing) { if (doForcing) presolveActions_ &= ~32; else presolveActions_ |= 32; } /// Whether we want to do impliedfree part of presolve inline bool doImpliedFree() const { return (presolveActions_ & 64) == 0; } inline void setDoImpliedFree(bool doImpliedfree) { if (doImpliedfree) presolveActions_ &= ~64; else presolveActions_ |= 64; } /// Whether we want to do dupcol part of presolve inline bool doDupcol() const { return (presolveActions_ & 128) == 0; } inline void setDoDupcol(bool doDupcol) { if (doDupcol) presolveActions_ &= ~128; else presolveActions_ |= 128; } /// Whether we want to do duprow part of presolve inline bool doDuprow() const { return (presolveActions_ & 256) == 0; } inline void setDoDuprow(bool doDuprow) { if (doDuprow) presolveActions_ &= ~256; else presolveActions_ |= 256; } /// Whether we want to do dependency part of presolve inline bool doDependency() const { return (presolveActions_ & 32768) != 0; } inline void setDoDependency(bool doDependency) { if (doDependency) presolveActions_ |= 32768; else presolveActions_ &= ~32768; } /// Whether we want to do singleton column part of presolve inline bool doSingletonColumn() const { return (presolveActions_ & 512) == 0; } inline void setDoSingletonColumn(bool doSingleton) { if (doSingleton) presolveActions_ &= ~512; else presolveActions_ |= 512; } /// Whether we want to do gubrow part of presolve inline bool doGubrow() const { return (presolveActions_ & 1024) == 0; } inline void setDoGubrow(bool doGubrow) { if (doGubrow) presolveActions_ &= ~1024; else presolveActions_ |= 1024; } /// Whether we want to do twoxtwo part of presolve inline bool doTwoxTwo() const { return (presolveActions_ & 2048) != 0; } inline void setDoTwoxtwo(bool doTwoxTwo) { if (!doTwoxTwo) presolveActions_ &= ~2048; else presolveActions_ |= 2048; } /// Whether we want to allow duplicate intersections inline bool doIntersection() const { return (presolveActions_ & 4096) != 0; } inline void setDoIntersection(bool doIntersection) { if (doIntersection) presolveActions_ &= ~4096; else presolveActions_ |= 4096; } /** How much we want to zero small values from aggregation - ratio 0 - 1.0e-12, 1 1.0e-11, 2 1.0e-10, 3 1.0e-9 */ inline int zeroSmall() const { return (presolveActions_&(8192|16384))>>13; } inline void setZeroSmall(int value) { presolveActions_ &= ~(8192|16384); presolveActions_ |= value<<13; } /// Set whole group inline int presolveActions() const { return presolveActions_ & 0xffff; } inline void setPresolveActions(int action) { presolveActions_ = (presolveActions_ & 0xffff0000) | (action & 0xffff); } /// Substitution level inline void setSubstitution(int value) { substitution_ = value; } /// Asks for statistics inline void statistics() { presolveActions_ |= 0x80000000; } /// Return presolve status (0,1,2) int presolveStatus() const; /**@name postsolve - postsolve the problem. If the problem has not been solved to optimality, there are no guarantees. If you are using an algorithm like simplex that has a concept of "basic" rows/cols, then set updateStatus Note that if you modified the original problem after presolving, then you must ``undo'' these modifications before calling postsolve. This version updates original*/ virtual void postsolve(bool updateStatus = true); /// Gets rid of presolve actions (e.g.when infeasible) void destroyPresolve(); /**@name private or protected data */ private: /// Original model - must not be destroyed before postsolve ClpSimplex * originalModel_; /// ClpPresolved model - up to user to destroy by deleteClpPresolvedModel ClpSimplex * presolvedModel_; /** "Magic" number. If this is non-zero then any elements with this value may change and so presolve is very limited in what can be done to the row and column. This is for non-linear problems. One could also allow for cases where sign of coefficient is known. */ double nonLinearValue_; /// Original column numbers int * originalColumn_; /// Original row numbers int * originalRow_; /// Row objective double * rowObjective_; /// The list of transformations applied. const CoinPresolveAction *paction_; /// The postsolved problem will expand back to its former size /// as postsolve transformations are applied. /// It is efficient to allocate data structures for the final size /// of the problem rather than expand them as needed. /// These fields give the size of the original problem. int ncols_; int nrows_; CoinBigIndex nelems_; /// Number of major passes int numberPasses_; /// Substitution level int substitution_; #ifndef CLP_NO_STD /// Name of saved model file std::string saveFile_; #endif /** Whether we want to skip dual part of presolve etc. 512 bit allows duplicate column processing on integer columns and dual stuff on integers */ int presolveActions_; protected: /// If you want to apply the individual presolve routines differently, /// or perhaps add your own to the mix, /// define a derived class and override this method virtual const CoinPresolveAction *presolve(CoinPresolveMatrix *prob); /// Postsolving is pretty generic; just apply the transformations /// in reverse order. /// You will probably only be interested in overriding this method /// if you want to add code to test for consistency /// while debugging new presolve techniques. virtual void postsolve(CoinPostsolveMatrix &prob); /** This is main part of Presolve */ virtual ClpSimplex * gutsOfPresolvedModel(ClpSimplex * originalModel, double feasibilityTolerance, bool keepIntegers, int numberPasses, bool dropNames, bool doRowObjective, const char * prohibitedRows=NULL, const char * prohibitedColumns=NULL); }; #endif