Tabu search method for combinatorial optimization supported with Wolfram Mathematica Software

Abstract

In this paper we present the results obtained from an algorithm based on the Tabu Search that was programmed using the commercial software Wolfram Mathematica. In Wolfram Language different implementations of random instances and others available in the library TSPLIB were made, comparing them later, with the results provided with the same algorithm, in the programming environment Visual Basic 6.0. The improvements that were obtained are due to the structuring of predesigned functions that allowed specifically analyzing two aspects: the optimization of the solution and its exploration in the neighborhoods where the presence of the optimum was already known. For this we focus on developing an oscillation in the tabu matrix analogously to what is applied to solutions where it is perceived that the global optimum could be. Finally, conclusive results are shown that allow observing the good performance of the program Wolfram Mathematica to deal with this type of problems, through the proper structuring of its internal functions.