Design-Optimization of Mechanical Components Using the “Multiple Objective Tabu Search” Method

The multiple objective version of the tabu search (TS algorithm was initially developed by Baykasoglu et al [1-3]. The idea of applying tabu search to multiple objective optimization, inspired from its solution structure, in which tabu search works with more than one solution (neighborhood solutions) at a time. This situation creates the opportunity to evaluate multiple objectives simultaneously in one run. To enable the original tabu search algorithm to work with more than one objective the selection and updating stages were redefined. Other stages are identical to the original tabu search algorithm. In this paper, mutiple objective tabu search algorithm is used to solve mechanical component design problems) with multiple objectives. Although there exists a number of classical techniques, meta-heuristic algorithms including TS have an edge over the classical methods in that they can find multiple Pareto-optimal solutions in one single run. In the paper four mechanical component design problems borrowed from the literature are solved. The results are compared with several other solution techniques including multiple objective genetic algorithms. It is observed that in many of the cases the multiple objective tabu search algorithm can find better and much wider spread of solutions than the compared algorithms.