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 | |
---|
22 | class ClpSimplexDual : public ClpSimplex { |
---|
23 | |
---|
24 | public: |
---|
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 | */ |
---|
115 | |
---|
116 | int dual(int ifValuesPass); |
---|
117 | /** For strong branching. On input lower and upper are new bounds |
---|
118 | while on output they are change in objective function values |
---|
119 | (>1.0e50 infeasible). |
---|
120 | Return code is 0 if nothing interesting, -1 if infeasible both |
---|
121 | ways and +1 if infeasible one way (check values to see which one(s)) |
---|
122 | Solutions are filled in as well - even down, odd up - also |
---|
123 | status and number of iterations |
---|
124 | */ |
---|
125 | int strongBranching(int numberVariables,const int * variables, |
---|
126 | double * newLower, double * newUpper, |
---|
127 | double ** outputSolution, |
---|
128 | int * outputStatus, int * outputIterations, |
---|
129 | bool stopOnFirstInfeasible=true, |
---|
130 | bool alwaysFinish=false); |
---|
131 | //@} |
---|
132 | |
---|
133 | /**@name Functions used in dual */ |
---|
134 | //@{ |
---|
135 | /** This has the flow between re-factorizations |
---|
136 | Broken out for clarity and will be used by strong branching |
---|
137 | |
---|
138 | Reasons to come out: |
---|
139 | -1 iterations etc |
---|
140 | -2 inaccuracy |
---|
141 | -3 slight inaccuracy (and done iterations) |
---|
142 | +0 looks optimal (might be unbounded - but we will investigate) |
---|
143 | +1 looks infeasible |
---|
144 | +3 max iterations |
---|
145 | |
---|
146 | If givenPi not NULL then in values pass |
---|
147 | */ |
---|
148 | int whileIterating(double * & givenPi); |
---|
149 | /** The duals are updated by the given arrays. |
---|
150 | Returns number of infeasibilities. |
---|
151 | After rowArray and columnArray will just have those which |
---|
152 | have been flipped. |
---|
153 | Variables may be flipped between bounds to stay dual feasible. |
---|
154 | The output vector has movement of primal |
---|
155 | solution (row length array) */ |
---|
156 | int updateDualsInDual(CoinIndexedVector * rowArray, |
---|
157 | CoinIndexedVector * columnArray, |
---|
158 | CoinIndexedVector * outputArray, |
---|
159 | double theta, |
---|
160 | double & objectiveChange, |
---|
161 | bool fullRecompute); |
---|
162 | /** The duals are updated by the given arrays. |
---|
163 | This is in values pass - so no changes to primal is made |
---|
164 | */ |
---|
165 | void updateDualsInValuesPass(CoinIndexedVector * rowArray, |
---|
166 | CoinIndexedVector * columnArray, |
---|
167 | double theta); |
---|
168 | /** While updateDualsInDual sees what effect is of flip |
---|
169 | this does actuall flipping. |
---|
170 | If change >0.0 then value in array >0.0 => from lower to upper |
---|
171 | */ |
---|
172 | void flipBounds(CoinIndexedVector * rowArray, |
---|
173 | CoinIndexedVector * columnArray, |
---|
174 | double change); |
---|
175 | /** |
---|
176 | Row array has row part of pivot row |
---|
177 | Column array has column part. |
---|
178 | This chooses pivot column. |
---|
179 | Spare arrays are used to save pivots which will go infeasible |
---|
180 | We will check for basic so spare array will never overflow. |
---|
181 | If necessary will modify costs |
---|
182 | For speed, we may need to go to a bucket approach when many |
---|
183 | variables are being flipped |
---|
184 | |
---|
185 | */ |
---|
186 | void dualColumn(CoinIndexedVector * rowArray, |
---|
187 | CoinIndexedVector * columnArray, |
---|
188 | CoinIndexedVector * spareArray, |
---|
189 | CoinIndexedVector * spareArray2, |
---|
190 | double accpetablePivot, |
---|
191 | CoinBigIndex * dubiousWeights); |
---|
192 | /** |
---|
193 | Row array has row part of pivot row |
---|
194 | Column array has column part. |
---|
195 | This sees what is best thing to do in dual values pass |
---|
196 | Returns 0 if theta_ move will put basic variable out to bound, |
---|
197 | 1 if can change dual on chosen row and leave variable in basis |
---|
198 | */ |
---|
199 | int checkPossibleValuesMove(CoinIndexedVector * rowArray, |
---|
200 | CoinIndexedVector * columnArray, |
---|
201 | double acceptablePivot, |
---|
202 | CoinBigIndex * dubiousWeights); |
---|
203 | /** |
---|
204 | This sees if we can move duals in dual values pass. |
---|
205 | This is done before any pivoting |
---|
206 | */ |
---|
207 | void doEasyOnesInValuesPass(double * givenReducedCosts); |
---|
208 | /** |
---|
209 | Chooses dual pivot row |
---|
210 | Would be faster with separate region to scan |
---|
211 | and will have this (with square of infeasibility) when steepest |
---|
212 | For easy problems we can just choose one of the first rows we look at |
---|
213 | |
---|
214 | If alreadyChosen >=0 then in values pass and that row has been |
---|
215 | selected |
---|
216 | */ |
---|
217 | void dualRow(int alreadyChosen); |
---|
218 | /** Checks if any fake bounds active - if so returns number and modifies |
---|
219 | updatedDualBound_ and everything. |
---|
220 | Free variables will be left as free |
---|
221 | Returns number of bounds changed if >=0 |
---|
222 | Returns -1 if not initialize and no effect |
---|
223 | Fills in changeVector which can be used to see if unbounded |
---|
224 | and cost of change vector |
---|
225 | */ |
---|
226 | int changeBounds(bool initialize,CoinIndexedVector * outputArray, |
---|
227 | double & changeCost); |
---|
228 | /** As changeBounds but just changes new bounds for a single variable. |
---|
229 | Returns true if change */ |
---|
230 | bool changeBound( int iSequence); |
---|
231 | /// Restores bound to original bound |
---|
232 | void originalBound(int iSequence); |
---|
233 | /** Checks if tentative optimal actually means unbounded in dual |
---|
234 | Returns -3 if not, 2 if is unbounded */ |
---|
235 | int checkUnbounded(CoinIndexedVector * ray,CoinIndexedVector * spare, |
---|
236 | double changeCost); |
---|
237 | /** Refactorizes if necessary |
---|
238 | Checks if finished. Updates status. |
---|
239 | lastCleaned refers to iteration at which some objective/feasibility |
---|
240 | cleaning too place. |
---|
241 | |
---|
242 | type - 0 initial so set up save arrays etc |
---|
243 | - 1 normal -if good update save |
---|
244 | - 2 restoring from saved |
---|
245 | */ |
---|
246 | void statusOfProblemInDual(int & lastCleaned, int type, |
---|
247 | ClpSimplexProgress * progress, |
---|
248 | double * givenDjs); |
---|
249 | /// Perturbs problem (method depends on perturbation()) |
---|
250 | void perturb(); |
---|
251 | /** Fast iterations. Misses out a lot of initialization. |
---|
252 | Normally stops on maximum iterations, first re-factorization |
---|
253 | or tentative optimum. If looks interesting then continues as |
---|
254 | normal. Returns 0 if finished properly, 1 otherwise. |
---|
255 | */ |
---|
256 | int fastDual(bool alwaysFinish=false); |
---|
257 | /** Checks number of variables at fake bounds. This is used by fastDual |
---|
258 | so can exit gracefully before end */ |
---|
259 | int numberAtFakeBound(); |
---|
260 | |
---|
261 | /** Pivot in a variable and choose an outgoing one. Assumes dual |
---|
262 | feasible - will not go through a reduced cost. Returns step length in theta |
---|
263 | Returns ray in ray_ (or NULL if no pivot) |
---|
264 | Return codes as before but -1 means no acceptable pivot |
---|
265 | */ |
---|
266 | int pivotResult(); |
---|
267 | |
---|
268 | //@} |
---|
269 | }; |
---|
270 | #endif |
---|