10 December 2022 in Computers and Electronics in Agriculture
21 December 2023 in IEEE Transactions on Evolutionary Computation
Both dual-population and two-phase strategies are effective for utilizing infeasible solution information and significantly enhancing the ability of algorithms to solve constrained multi-objective optimization problems. However, most existing algorithms tend to underperform when facing problems with complex constraints. To address these issues, a constrained multi-objective evolutionary algorithm named DPTPEA, which combines dual-population and two-phase strategies, is proposed in this paper. DPTPEA employs two collaborative populations (the exploitive population and the tractive population) and divides the evolutionary process of the tractive population into two phases (Phase 1 and Phase 2). In Phase 1, the tractive population ignores constraints and drags the exploitive population across the infeasible region by sharing offspring information. In Phase 2, the tractive population adopts the epsilon-constrained method to converge toward the constrained Pareto front and to guide the exploitive population exploiting different feasible regions. Moreover, a dynamic cooperation strategy, a boundary point direction sampling strategy, and a dynamic environmental selection are proposed to improve the exploration ability of tractive population for solving complex problems. Comprehensive experiments on three popular test suites demonstrate that DPTPEA outperforms seven state-of-the-art algorithms on most test problems.
28 August 2022 in Neural Computing and Applications
25 January 2022 in IEEE Transactions on Evolutionary Computation
Multiobjective evolutionary algorithms (EAs) based on decomposition are becoming successful and popular. Particularly, a conical area EA (CAEA) was developed to heighten the convergence and population diversity of decomposition-based algorithms for bi-objective optimization by employing a cone decomposition approach. The global Pareto optimality of cone decomposition was proved. It means that the optimum of every conical subproblem of the cone decomposition approach is Pareto optimal in the entire bi-objective space in the presence of a continuous frontier segment within its associated subregion. In this article, based on the global Pareto optimality, we further reveal that the cone decomposition approach also has a desired ability of $\mu $ -distribution convergence for bi-objective optimization. We first prove by the monotonicity and additivity of the Lebesgue measure and the squeeze theorem for limits that the hypervolume of the optimal solutions of all $\mu $ subproblems of the cone decomposition approach is infinitely close to that of the optimal $\mu $ -distribution for large enough $\mu $ in the presence of a continuous frontier. Meanwhile, the original CAEA is further improved to better approximate the optimal solutions of conical subproblems. The experimental results on bi-objective benchmark problems show that the improved CAEA achieves higher qualities of frontiers and better $\mu $ -distribution convergence in the terms of hypervolume error compared with six other popular multiobjective algorithms.