Orbital Shrinking: A New Tool for Hybrid MIP/CP Methods
Orbital shrinking is a newly developed technique in the MIP community to deal with symmetry issues, which is based on aggregation rather than on symmetry breaking. In a recent work, a hybrid MIP/CP scheme based on orbital shrinking was developed for the multi-activity shift scheduling problem, showing significant improvements over previous pure MIP approaches. In the present paper we show that the scheme above can be extended to a general framework for solving arbitrary symmetric MIP instances. This framework naturally provides a new way for devising hybrid MIP/CP decompositions. Finally, we specialize the above framework to the multiple knapsack problem. Computational results show that the resulting method can be orders of magnitude faster than pure MIP approaches on hard symmetric instances.
KeywordsKnapsack Problem Master Problem Steiner Triple System Multiple Knapsack Problem Symmetry Breaking Constraint
Unable to display preview. Download preview PDF.
- 2.Gent, I.P., Petrie, K.E., Puget, J.F.: Symmetry in constraint programming. In: Rossi, F., van Beek, P., Walsh, T. (eds.) Handbook of Constraint Programming, pp. 329–376. Elsevier (2006)Google Scholar
- 6.McKay, B.D.: Practical graph isomorphism (1981)Google Scholar
- 10.Martello, S., Toth, P.: Knapsack Problems: Algorithms and Computer Implementations. Wiley (1990)Google Scholar
- 12.Gent, I.P., Walsh, T.: From approximate to optimal solutions: Constructing pruning and propagation rules. In: IJCAI, pp. 1396–1401. Morgan Kaufmann (1997)Google Scholar
- 14.IBM ILOG: CPLEX 12.4 User’s Manual (2011)Google Scholar
- 15.Gecode Team: Gecode: Generic constraint development environment (2012), http://www.gecode.org