Scippy

SCIP

Solving Constraint Integer Programs

xternal_binpacking.c
Go to the documentation of this file.
1 /* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
2 /* */
3 /* This file is part of the program and library */
4 /* SCIP --- Solving Constraint Integer Programs */
5 /* */
6 /* Copyright (C) 2002-2017 Konrad-Zuse-Zentrum */
7 /* fuer Informationstechnik Berlin */
8 /* */
9 /* SCIP is distributed under the terms of the ZIB Academic License. */
10 /* */
11 /* You should have received a copy of the ZIB Academic License */
12 /* along with SCIP; see the file COPYING. If not email to scip@zib.de. */
13 /* */
14 /* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
15 
16 /**@file xternal_binpacking.c
17  * @brief The bin packing example of SCIP
18  * @author Timo Berthold
19  * @author Stefan Heinz
20  */
21 
22 /*--+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0----+----1----+----2*/
23 
24 /**@page BINPACKING_MAIN Overview
25  * @author Timo Berthold
26  * @author Stefan Heinz
27  *
28  * This example contains a branch-and-price approach for the binpacking problem which is realized with the framework
29  * \SCIP. Therefore, the following plugins are implemented:
30  *
31  * - a \ref reader_bpa.c "problem reader" which parses the problem out of file and creates the corresponding problem within \SCIP
32  * - a \ref probdata_binpacking.c "(global) problem data structure" which contains all necessary information
33  * - a \ref pricer_binpacking.c "pricer" which generates new variables/columns during the search
34  * - the \ref branch_ryanfoster.c "Ryan/Foster branching rule"
35  * - a \ref cons_samediff.c "constraint handler" which handles the branching decisions of the Ryan/Foster branching
36  * - a \ref vardata_binpacking.c "variable data structure" which stores information for each variable and is needed to perform the Ryan/Foster
37  * branching
38  *
39  * In the following we introduce the problem, explain the use of the reader plugin and pricer plugin. Finally, we
40  * introduce the Ryan/Foster branching rule and briefly discuss how that specific branching rule is realized within
41  * the framework \SCIP.
42  *
43  * -# \ref BINPACKING_PROBLEM "Problem description"
44  * -# \ref BINPACKING_READER "Parsing the input format and creating the problem"
45  * -# \ref BINPACKING_PROBLEMDATA "Main problem data"
46  * -# \ref BINPACKING_PRICER "Pricing new variables"
47  * -# \ref BINPACKING_BRANCHING "Ryan/Foster branching"
48  * -# \ref BINPACKING_MAKEFILE "The Makefile"
49  *
50  */
51 
52 /**@page BINPACKING_PROBLEM Problem description
53  *
54  * The binpacking problem consists of the task to distribute a given set of items \f$ [n] := \{1, \dots, n\}\f$ with
55  * nonnegative size \f$s_i\f$ to a minimal number of bins, all of the same capacity \f$\kappa\f$. As example consider 9
56  * items with sizes: 2, 1, 2, 1, 1, 2, 3, 2, and 1 and a bin capacity of \f$\kappa\f$ of 4. The following pictures show
57  * a feasible solution which needs 5 bins. The minimum number of bins needed for that example is 3.
58  *
59  * <CENTER>
60  * \image html binpacking.png
61  * </CENTER>
62  *
63  * This problem can be formulated as a set covering problem as discussed by Gilmore and Gomory in the following two classical papers:
64  * - Gilmore P. C., R. E. Gomory: A linear programming approach to the cutting-stock problem. Operations Research 9 (1961): 849-859.
65  * - Gilmore P. C., R. E. Gomory: A linear programming approach to the cutting-stock problem - Part II. Operations Research 11 (1963): 863-888
66  *
67  * We introduce a binary variable \f$x_{S}\f$ for each feasible packing \f$S\f$. A <b>packing</b> \f$S\f$ is an
68  * assignment vector \f$ \lambda_{S}\in\{0,1\}^n \f$ which states the items belonging to that packing. It is
69  * <b>feasible</b>, if and only if the total size of the items contained in this assignment is not greater than the
70  * given capacity \f$\kappa\f$. Let \f$\mathcal{S}\f$ be the set of all feasible packing, this measns:
71  *
72  * \f[
73  * \mathcal{S} := \{S\subseteq [n] \mid \sum_{i:i\in S} s_{i} \leq \kappa \}
74  * \f]
75  *
76  * An integer program can be formulated as follows:
77  *
78  * \f[
79  * \begin{array}[t]{rll}
80  * \min & \displaystyle \sum_{S \in \mathcal{S}} x_{S} \\
81  * & \\
82  * subject \ to & \displaystyle \sum_{S \in \mathcal{S}} (\lambda_{S})_{i}x_{S} \ge 1 & \quad \forall i \in \{1,\dots,n\} \\
83  * & \\
84  * & x_{S} \in \{0,1\} & \quad \forall S \in \mathcal{S} \\
85  * \end{array}
86  * \f]
87  *
88  * This means we are searching for a set of packings such that each item is contained in at least one of the selected
89  * packings. Since the objective is to minimize the number of used packings, each optimal solution to the above problem
90  * can be transformed into a solution where each item is packed exactly once with the same cost.
91  *
92  *
93  * Since \f$\mathcal{S}\f$ can be of exponential size, we are using a column generation approach to solve this
94  * problem. We initialize the (master) problem with a set of \f$ n \f$ variables representing packings of a single item
95  * per bin. Now, we have to iteratively search for variables representing "better" packings, i.e., a packing pattern
96  * which reduces the overall cost. For a given solution \f$y^\star\f$ of the (restricted) dual linear program, we have
97  * to find a variable/packing \f$ \lambda_{S} \f$ for which the reduced costs is negative. This means:
98  *
99  * \f[
100  * c_{S} - \sum_{i=0}^n (\lambda_S)_i y_i^\star < 0.
101  * \f]
102  *
103  * Since all variables \f$ \lambda_{S} \f$ have an objective coefficient \f$ c_{S} = 1 \f$ the above condition is
104  * equivalent to
105  *
106  * \f[
107  * \sum_{i=0}^n (\lambda_S)_i y_i^\star > 1.
108  * \f]
109  *
110  *
111  * To find such a variable/packing we solve the following integer program:
112  *
113  * \f[
114  * \begin{array}[t]{rll}
115  * \max & \displaystyle \sum_{i=1}^n (\lambda_S)_i y^\star_i\\
116  * & \\
117  * subject \ to & \displaystyle \sum_{i=0}^n (\lambda_S)_i s_i \leq \kappa \\
118  * & \\
119  * & (\lambda_S)_i \in \{0,1\} & \quad \forall i \in \{ 1, \dots , n \} \\
120  * \end{array}
121  * \f]
122  *
123  * where \f$ (\lambda_S)_i \f$ for \f$i\in\{1,\dots,n\}\f$ are binary variables and \f$y^\star_i\f$ given by the dual
124  * solution of the restricted master problem.
125  *
126  * The above problem is a knapsack problem which can be solved via dynamic programming or by solving the above integer
127  * program. In this example we implemented a pricer which solves the integer program.
128  *
129  */
130 
131 
132 
133 /**@page BINPACKING_MAKEFILE The Makefile
134  *
135  * The Makefile is based on the main \SCIP Makefile. This means that all compiling options which are
136  * available for \SCIP are also available for the binpacking project. Below, you find a list
137  * of the most important compiling flags, the values they can take, and a short description. The
138  * values in bold face are the default values.
139  *
140  * - <code>LPS={clp | cpx | none | <b>spx</b>}</code>
141  * <br>
142  * Defines the linear program solver to use:
143  * - <code>clp</code> use COIN-OR CLP
144  * - <code>cpx</code> use IBM CPLEX
145  * - <code>none</code> no LP solver
146  * - <code><b>spx</b></code> use SoPlex
147  *
148  * - <code>OPT={dbg | <b>opt</b>}</code>
149  * <br>
150  * Defines if the projects gets compiled in debug (<code>dbg</code>) mode or
151  * optimized (<code><b>opt</b></code>) mode. In the debug mode all assertions are checked.
152  *
153  * - <code>ZIMPL={false | <b>true</b>}</code>
154  * <br>
155  * Defines if the modeling language ZIMPL should be linked to binary or not.
156  *
157  * In the following we explain the all <b>Makefile targets</b>.
158  *
159  * - <b>lint</b>
160  * <br>
161  * Statically checks the code for uninitialized variables and many other possible problems,
162  * which even do not lead to compiler errors. For this,
163  * the external tool flexelint is needed. The call produces the file <code>lint.out</code>
164  * which contains all the detected warnings. From the experience when developing \SCIP, we strongly
165  * recommend to use such a code checker. It is always a surprising the stuff such tools detect.
166  * <br>
167  *
168  * - <b>clean</b>
169  * <br>
170  * Remove all objective files, libraries, and binaries.
171  * <br>
172  *
173  * - <b>test</b>
174  * <br>
175  * Starts an automated test run based on the SCIP test runs (see \ref TEST "How to run automated tests with SCIP").
176  * <br>
177  *
178  * - <b>tags</b>
179  * <br>
180  * Generates tags which can be used in the editor <b>emacs</b> and <b>xemacs</b>.
181  * <br>
182  *
183  * - <b>depend</b>
184  * <br>
185  * Generates the dependencies for the compiling process.
186  */