Constrained Multi-Objective Water Evaporation Optimization Algorithm Based on Decomposition With ε-Constraint Handling Technology

In order to solve the constrained multi-objective problem more effectively, the constrained multi-objective water evaporation optimization algorithm based on decomposition with ε-constraint handling technology (MOEA/D-WEO-εCHT) is proposed in this paper. MOEA/D-WEO-εCHT uses the MOEA/D framework to deal with multi-objective, ε-constraint handling technology to handle constraints and WEO algorithm to perform the main body evolution operation. Firstly, an external memory population in which the abandoned excellent parent can be kept is established for each subproblem, which make the abandoned parent be fully utilized by participating in the subsequent evolution. Secondly, some strategies of WEO, such as calculation and normalization method of individual fitness and calculation method of step size S, are improved to make WEO suitable for solving constrained multi-objective problems and helpful to obtain the Pareto solution set with better convergence and distribution. Thirdly, the ε(t) calculation method and retention mechanism of individuals are improved to reasonably coordinate feasible and infeasible individuals in different evolution stages. To evaluate the performance of the proposed algorithm, the CTP and CF benchmark instances and an engineering optimization problem are studied. The experimental results illustrate that MOEA/D-WEO-εCHT is more effective than other related algorithms on these benchmark instances, in terms of convergence and distribution of the obtained solution sets.