source: trunk/Clp/src/ClpSimplexDual.hpp @ 754

Last change on this file since 754 was 754, checked in by andreasw, 14 years ago

first version

  • Property svn:eol-style set to native
  • Property svn:keywords set to Author Date Id Revision
File size: 11.0 KB
Line 
1// Copyright (C) 2002, International Business Machines
2// Corporation and others.  All Rights Reserved.
3
4/*
5   Authors
6   
7   John Forrest
8
9 */
10#ifndef ClpSimplexDual_H
11#define ClpSimplexDual_H
12
13#include "ClpSimplex.hpp"
14
15/** This solves LPs using the dual simplex method
16
17    It inherits from ClpSimplex.  It has no data of its own and
18    is never created - only cast from a ClpSimplex object at algorithm time.
19
20*/
21
22class ClpSimplexDual : public ClpSimplex {
23
24public:
25
26  /**@name Description of algorithm */
27  //@{
28  /** Dual algorithm
29
30      Method
31
32     It tries to be a single phase approach with a weight of 1.0 being
33     given to getting optimal and a weight of updatedDualBound_ being
34     given to getting dual feasible.  In this version I have used the
35     idea that this weight can be thought of as a fake bound.  If the
36     distance between the lower and upper bounds on a variable is less
37     than the feasibility weight then we are always better off flipping
38     to other bound to make dual feasible.  If the distance is greater
39     then we make up a fake bound updatedDualBound_ away from one bound.
40     If we end up optimal or primal infeasible, we check to see if
41     bounds okay.  If so we have finished, if not we increase updatedDualBound_
42     and continue (after checking if unbounded). I am undecided about
43     free variables - there is coding but I am not sure about it.  At
44     present I put them in basis anyway.
45
46     The code is designed to take advantage of sparsity so arrays are
47     seldom zeroed out from scratch or gone over in their entirety.
48     The only exception is a full scan to find outgoing variable for
49     Dantzig row choice.  For steepest edge we keep an updated list
50     of infeasibilities (actually squares). 
51     On easy problems we don't need full scan - just
52     pick first reasonable.
53
54     One problem is how to tackle degeneracy and accuracy.  At present
55     I am using the modification of costs which I put in OSL and some
56     of what I think is the dual analog of Gill et al.
57     I am still not sure of the exact details.
58
59     The flow of dual is three while loops as follows:
60
61     while (not finished) {
62
63       while (not clean solution) {
64
65          Factorize and/or clean up solution by flipping variables so
66          dual feasible.  If looks finished check fake dual bounds.
67          Repeat until status is iterating (-1) or finished (0,1,2)
68
69       }
70
71       while (status==-1) {
72
73         Iterate until no pivot in or out or time to re-factorize.
74
75         Flow is:
76
77         choose pivot row (outgoing variable).  if none then
78         we are primal feasible so looks as if done but we need to
79         break and check bounds etc.
80
81         Get pivot row in tableau
82
83         Choose incoming column.  If we don't find one then we look
84         primal infeasible so break and check bounds etc.  (Also the
85         pivot tolerance is larger after any iterations so that may be
86         reason)
87
88         If we do find incoming column, we may have to adjust costs to
89         keep going forwards (anti-degeneracy).  Check pivot will be stable
90         and if unstable throw away iteration and break to re-factorize.
91         If minor error re-factorize after iteration.
92
93         Update everything (this may involve flipping variables to stay
94         dual feasible.
95
96       }
97
98     }
99
100     TODO's (or maybe not)
101
102     At present we never check we are going forwards.  I overdid that in
103     OSL so will try and make a last resort.
104
105     Needs partial scan pivot out option.
106
107     May need other anti-degeneracy measures, especially if we try and use
108     loose tolerances as a way to solve in fewer iterations.
109
110     I like idea of dynamic scaling.  This gives opportunity to decouple
111     different implications of scaling for accuracy, iteration count and
112     feasibility tolerance.
113
114     for use of exotic parameter startFinishoptions see Clpsimplex.hpp
115  */
116
117  int dual(int ifValuesPass,int startFinishOptions=0);
118  /** For strong branching.  On input lower and upper are new bounds
119      while on output they are change in objective function values
120      (>1.0e50 infeasible).
121      Return code is 0 if nothing interesting, -1 if infeasible both
122      ways and +1 if infeasible one way (check values to see which one(s))
123      Solutions are filled in as well - even down, odd up - also
124      status and number of iterations
125  */
126  int strongBranching(int numberVariables,const int * variables,
127                      double * newLower, double * newUpper,
128                      double ** outputSolution,
129                      int * outputStatus, int * outputIterations,
130                      bool stopOnFirstInfeasible=true,
131                      bool alwaysFinish=false,
132                      int startFinishOptions=0);
133  /// This does first part of StrongBranching
134  ClpFactorization * setupForStrongBranching(char * arrays, int numberRows, int numberColumns);
135  /// This cleans up after strong branching
136  void cleanupAfterStrongBranching();
137  //@}
138
139  /**@name Functions used in dual */
140  //@{
141  /** This has the flow between re-factorizations
142      Broken out for clarity and will be used by strong branching
143
144      Reasons to come out:
145      -1 iterations etc
146      -2 inaccuracy
147      -3 slight inaccuracy (and done iterations)
148      +0 looks optimal (might be unbounded - but we will investigate)
149      +1 looks infeasible
150      +3 max iterations
151
152      If givenPi not NULL then in values pass
153   */
154  int whileIterating(double * & givenPi,int ifValuesPass); 
155  /** The duals are updated by the given arrays.
156      Returns number of infeasibilities.
157      After rowArray and columnArray will just have those which
158      have been flipped.
159      Variables may be flipped between bounds to stay dual feasible.
160      The output vector has movement of primal
161      solution (row length array) */
162  int updateDualsInDual(CoinIndexedVector * rowArray,
163                  CoinIndexedVector * columnArray,
164                  CoinIndexedVector * outputArray,
165                  double theta,
166                  double & objectiveChange,
167                        bool fullRecompute);
168  /** The duals are updated by the given arrays.
169      This is in values pass - so no changes to primal is made
170  */
171  void updateDualsInValuesPass(CoinIndexedVector * rowArray,
172                  CoinIndexedVector * columnArray,
173                  double theta);
174  /** While updateDualsInDual sees what effect is of flip
175      this does actuall flipping.
176      If change >0.0 then value in array >0.0 => from lower to upper
177  */
178  void flipBounds(CoinIndexedVector * rowArray,
179                  CoinIndexedVector * columnArray,
180                  double change);
181  /**
182      Row array has row part of pivot row
183      Column array has column part.
184      This chooses pivot column.
185      Spare arrays are used to save pivots which will go infeasible
186      We will check for basic so spare array will never overflow.
187      If necessary will modify costs
188      For speed, we may need to go to a bucket approach when many
189      variables are being flipped.
190      Returns best possible pivot value
191  */
192  double dualColumn(CoinIndexedVector * rowArray,
193                  CoinIndexedVector * columnArray,
194                  CoinIndexedVector * spareArray,
195                  CoinIndexedVector * spareArray2,
196                  double accpetablePivot,
197                  CoinBigIndex * dubiousWeights);
198  /// Does first bit of dualColumn
199  int dualColumn0(const CoinIndexedVector * rowArray,
200                  const CoinIndexedVector * columnArray,
201                  CoinIndexedVector * spareArray,
202                  double acceptablePivot,
203                  double & upperReturn, double &bestReturn);
204  /**
205      Row array has row part of pivot row
206      Column array has column part.
207      This sees what is best thing to do in dual values pass
208      if sequenceIn==sequenceOut can change dual on chosen row and leave variable in basis
209  */
210  void checkPossibleValuesMove(CoinIndexedVector * rowArray,
211                               CoinIndexedVector * columnArray,
212                              double acceptablePivot);
213  /**
214      Row array has row part of pivot row
215      Column array has column part.
216      This sees what is best thing to do in branch and bound cleanup
217      If sequenceIn_ < 0 then can't do anything
218  */
219  void checkPossibleCleanup(CoinIndexedVector * rowArray,
220                               CoinIndexedVector * columnArray,
221                              double acceptablePivot);
222  /**
223      This sees if we can move duals in dual values pass.
224      This is done before any pivoting
225  */
226  void doEasyOnesInValuesPass(double * givenReducedCosts);
227  /**
228      Chooses dual pivot row
229      Would be faster with separate region to scan
230      and will have this (with square of infeasibility) when steepest
231      For easy problems we can just choose one of the first rows we look at
232     
233      If alreadyChosen >=0 then in values pass and that row has been
234      selected
235  */
236  void dualRow(int alreadyChosen);
237  /** Checks if any fake bounds active - if so returns number and modifies
238      updatedDualBound_ and everything.
239      Free variables will be left as free
240      Returns number of bounds changed if >=0
241      Returns -1 if not initialize and no effect
242      Fills in changeVector which can be used to see if unbounded
243      and cost of change vector
244  */
245  int changeBounds(bool initialize,CoinIndexedVector * outputArray,
246                   double & changeCost);
247  /** As changeBounds but just changes new bounds for a single variable.
248      Returns true if change */
249  bool changeBound( int iSequence);
250  /// Restores bound to original bound
251  void originalBound(int iSequence);
252  /** Checks if tentative optimal actually means unbounded in dual
253      Returns -3 if not, 2 if is unbounded */
254  int checkUnbounded(CoinIndexedVector * ray,CoinIndexedVector * spare,
255                     double changeCost);
256  /**  Refactorizes if necessary
257       Checks if finished.  Updates status.
258       lastCleaned refers to iteration at which some objective/feasibility
259       cleaning too place.
260
261       type - 0 initial so set up save arrays etc
262            - 1 normal -if good update save
263            - 2 restoring from saved
264  */
265  void statusOfProblemInDual(int & lastCleaned, int type,
266                             double * givenDjs, ClpDataSave & saveData,
267                             int ifValuesPass);
268  /// Perturbs problem (method depends on perturbation())
269  void perturb();
270  /** Fast iterations.  Misses out a lot of initialization.
271      Normally stops on maximum iterations, first re-factorization
272      or tentative optimum.  If looks interesting then continues as
273      normal.  Returns 0 if finished properly, 1 otherwise.
274  */
275  int fastDual(bool alwaysFinish=false);
276  /** Checks number of variables at fake bounds.  This is used by fastDual
277      so can exit gracefully before end */
278  int numberAtFakeBound();
279
280  /** Pivot in a variable and choose an outgoing one.  Assumes dual
281      feasible - will not go through a reduced cost.  Returns step length in theta
282      Returns ray in ray_ (or NULL if no pivot)
283      Return codes as before but -1 means no acceptable pivot
284  */
285  int pivotResult();
286  /** Get next free , -1 if none */
287  int nextSuperBasic();
288  /** Startup part of dual (may be extended to other algorithms)
289      returns 0 if good, 1 if bad */
290  int startupSolve(int ifValuesPass,double * saveDuals,int startFinishOptions);
291  void finishSolve(int startFinishOptions);
292  void gutsOfDual(int ifValuesPass,double * & saveDuals,int initialStatus,
293                  ClpDataSave & saveData);
294  //int dual2(int ifValuesPass,int startFinishOptions=0);
295 
296  //@}
297};
298#endif
Note: See TracBrowser for help on using the repository browser.