Branch and Bound Algorithm Selection by Performance Prediction

Wepropose a method called Selection by Performance Prediction (SPP) which allows one, when faced with a particular problem instance, to select a Branch and Boundalgorithm from amongseveral promising ones. This method is based on Knuth’s sampling method which estimates the efficiency of a backtrack program on a particular instance by iteratively generating randompaths in the search tree. Wepresent a simple adaptation of this estimator in the field of combinatorial optimization problems, more precisely for an extension of the maximalconstraint satisfaction framework. Experiments both on randomand strongly structured instances showthat, in most cases, the proposed methodis able to select, froma candidate list, the best algorithm for solving a given instance.

Authors

• Lionel Lobjois Michel Lemaître

Keywords

No keywords are indexed for this paper.