Ensemble for Solving Quadratic Assignment Problems
L. Q. Song, M.H. Lim, Ponnuthurai Nagaratnam Suganthan, Van Khanh Doan · 2009
In this paper, we present a scheme whereby diverse optimization algorithms are incorporated within a framework of selective reproduction according to fitness. By forming an ensemble of several populated optimization algorithms, it is shown that the exploitative traits can be extended across several search algorithms. Results of simulations on several difficult quadratic assignment problem benchmarks based on a fixed computational time budget have shown that the ensemble scheme convincingly outperforms the individual constituent optimization algorithms.