A multiple-search multi-start framework for metaheuristics
Chun‐Wei Tsai, Kai-Cheng Hu, Ming‐Chao Chiang · 2014
Until now, most, if not all, of the metaheuristic algorithms have been extremely sensitive to the initial solutions and may even converge to a local optimum at early iterations for most optimization problems. This paper introduces an effective and efficient framework, called multiple-search multi-start (MSMS), to mitigate the impact of these problems. To evaluate the performance of the proposed framework, we apply it to k-means and particle swarm optimization for the clustering problem and compare the results with those of several well-known clustering algorithms. The experimental results show that the proposed framework can significantly enhance the performance of not only single-solution-based but also population-based metaheuristic algorithms in terms of both the quality and the computation time.