A C++ implementation of a skeleton for Tabu Search method

Abstract

We present a C++ implementation of a skeleton for Tabu Search method. Tabu Search method is a well-known meta-heuristic that has proved successful for sub-optimally solving hard combinatorial optimization problems. This method has been already implemented for a large number of optimization problems, to the best of our knowledge they are all ad hoc implementations.

We propose a generic C++ implementation based on a skeleton design for the method. This implementation offers, among others, the possibility for the user to instantiate the Tabu Search method for any problem with little efforts and basic knowledge of C++ language. The implementation provides both robustness and re-usability properties. We have instantiated the skeleton on three different problems, namely the Quadratic Assignment, Minimum k-Cardinality Tree and 0-1 Multi-dimensional Knapsack problems so as to evidence the claimed properties of the skeleton.

Postprint (published version)