A Fast Strength Pareto Evolutionary Algorithm Incorporating Predefined Preference Information

Strength Pareto Evolutionary Algorithm 2 (SPEA2) has achieved great success for handling multiobjective optimization problems. However, it has been widely reported that SPEA2 gets subjected to a huge amount of computational effort while pursuing a good distribution of approximated solutions. This paper explores a new way to keep the good properties of SPEA2 and reduce its high computational burden simultaneously, with the aid of predefined preference information. By incorporating preference information, the proposed fast SPEA (FSPEA) can efficiently perform individuals' density estimation and environmental selection, thus speeding up the whole running time of the evolution process. Empirical studies show that the proposed FSPEA algorithm can obtain very competitive performance on a number of multiobjective test problems considered in this paper.