This paper shows that the many greedy strategies that have been designed to repair infeasible solutions to multi-objective knapsack problems (MOKPs) with small item differences perform poorly when item differences are large. To effectively solve different types of MOKPs, this paper proposes a greedy strategy to improve the quality of feasible and infeasible solutions. It repairs all of the infeasible solutions to feasible solutions, and then maximizes the quality of each feasible solution under the limitations of knapsack capacities. Simulation experiments on different types of MOKPs show that the proposed strategy is superior to existing strategies. Compared with MOGLS, MOEA/D, and MOEA/D-M2M, the proposed evolutionary framework performs better in solving different MOKPs.

本文表明，当项目差异很大时，许多设计用于修复具有小项目差异的多目标背包问题 (MOKP) 的不可行解决方案的贪婪策略表现不佳。为了有效解决不同类型的MOKP，本文提出了一种贪婪策略来提高可行解和不可行解的质量。它将所有不可行解修复为可行解，然后在背包容量的限制下最大化每个可行解的质量。对不同类型 MOKP 的仿真实验表明，所提出的策略优于现有策略。与 MOGLS、MOEA/D 和 MOEA/D-M2M 相比，所提出的进化框架在解决不同的 MOKPs 方面表现更好。