Abstract: Aiming to consider the stochastic maintenance environment and the diverse decision criteria,and further enhance the practicality of traditional sequential maintenance strategies,a multi-objective phased sequential maintenance strategy under stochasticity was studied.Accordingly,a multi-objective stochastic programming model with objectives of minimizing the total maintenance cost rate and maximizing the availability subject to reliability constraints was formulated.According to the stochastic and multi-objective natures of the configured model,a multi-objective black-winged kite algorithm combing a stochastic simulation method was developed to cope with it.In the designed approach,the multi-objective black winged kite algorithm and the stochastic simulation method were severally utilized to find candidate solutions and evaluate their fitness values under stochasticity.By taking a key part of a ship as an example,the devised approach was compared with multi-objective non-dominated sorting genetic algorithm-Ⅱ,multi-objective evolutionary algorithm based on decomposition and multi-objective particle swarm optimization,and the experimental results demonstrated the feasibility and competitiveness of the proposed model and method in tackling the concerned problem.

Key words: multi-objective phased sequential maintenance, reliability, availability, stochastic simulation method, multi-objective black-winged kite algorithm

摘要： 基于考虑维修环境的随机性、决策目标的多样性以及提高传统顺序维修策略实操性的现实需求,提出随机环境下的多目标分阶段顺序维修策略,建立以维修费用率最小化与可用度最大化为目标,以可靠度为约束的多目标随机规划数学模型。根据所建立模型的随机、多目标特性,设计了一种结合了随机仿真方法的多目标黑翅鸢优化算法进行求解。多目标黑翅鸢优化算法和随机仿真方法分别用于搜索候选解和在随机环境下评估解的适应度值。以某船舶关键部件为例,将所设计方法与非支配排序遗传算法Ⅱ、基于分解的多目标进化算法和多目标粒子群优化算法进行对比分析,实验结果验证了所提模型与算法在解决该问题上的可行性和高效性。

关键词: 多目标分阶段顺序维修, 可靠性, 可用度, 随机仿真方法, 多目标黑翅鸢优化算法