A Reference-Vector-Based Strength Pareto Evolutionary Algorithm 2

Abstract

In this paper, a reference-vector-based strength Pareto evolutionary algorithm 2 (RVSPEA2) is proposed to deal with the multiobjective continuous optimization problems. In the proposed RVSPEA2, an objective normalization technique is firstly applied to guarantee the consistency of disparately scaled objectives. Then an improved solutions selection mechanism, based on the reference vectors generation and niche-selection operation, is designed to improve the diversity and convergence of the optimal solutions. Finally, some benchmark test problems are applied to evaluate the effectiveness of the proposed RVSPEA2 algorithm. The results showed that this algorithm performs well than other compared optimization algorithms on convergence and diversity.