Search

To address constrained optimization problems (COPs), there are two crucial trade-offs in preeminent constrained evolutionary algorithms: (1) the trade-off between exploration and exploitation, and (2) the one between the constraints and the objectives. Moreover, the problem of inaccurate diversity regulation in constrained optimization evolutionary algorithms has yet to be completely solved. To achieve the first trade-off, this paper designs a diversity controller (DC) based on a small-world network. Then, the fitness distance correlation information is applied to adjust the reconnection probability of the small-world network to dynamically control the diversity. To achieve the second trade-off, we propose an infeasible–feasible regions constraint handling method (IF). There are two stages in the IF method. The first stage is to search the boundary between the infeasible and feasible regions, while the second stage adopts the self-adaptive Epsilon constraint handling method. Furthermore, combining the DC and IF constraint handing methods into the success-history-based parameter adaptive differential evolution (SHADE) algorithm, a DC-SHADE-IF algorithm is proposed. Twenty-eight constrained optimization problems provided in the CEC2017 competition on constrained real parameter optimization are employed to compare the performance between the proposed DC-SHADE-IF and seven state-of-the-art algorithms. Besides that, two real-world constrained optimization problems are employed to examine the performance of the proposed DC-SHADE-IF in real-world problems. Experimental results show the superior performance of the proposed algorithm in terms of accuracy and convergence. • A diversity controller is proposed to control population diversity. • An infeasible–feasible regions constraint handling method. • Experiments were outperform in CEC 2017 and real-world problems. [ABSTRACT FROM AUTHOR]