Abstract: To solve the problem of service composition optimization,which is affected by multiple manufacturing tasks,cross-region manufacturing services and service capability constraints,a multi-objective optimization method for manufacturing service composition with service capability constraints was proposed.By considering the constraints such as the decomposition of heterogeneous tasks,the vertical execution order and horizontal processing order of atomic tasks,the cross-regional distribution of manufacturing services and the services capacity,a multi-objective optimization model to minimize the maximum completion time and the total cost was formulated.Aiming at the characteristics of this problem,a Heuristic Search based  Non-dominated Sorting Genetic AlgorithmⅡ(HSNSGA-Ⅱ) was proposed.The heuristic search was applied in the initial population,crossover and mutation stages to improve the search quality.The effectiveness of HSNSGA-Ⅱ was verified by comparison with three heuristic algorithms,and the practicability of HSNSGA-Ⅱ was further verified by application case analysis.

Key words: manufacturing service composition, multi-objective optimization, service capability constraints, non-dominated sorting genetic algorithm Ⅱ, heuristic search

摘要： 为了解决受制造任务种类多、制造服务跨地域和服务能力受限等特点影响的服务组合优化问题,提出服务能力受限下制造服务组合的多目标优化方法。该方法考虑异构任务的分解、原子任务的纵向执行顺序和横向处理顺序、制造服务跨地域分布和能力受限因素等约束,以最小化最大完成时间和最小化总成本为目标构建多目标优化模型。针对该问题的特点,提出基于启发式搜索的非支配排序算法(HSNSGA-Ⅱ),在初始化种群、交叉和变异阶段引入启发式搜索机制,能够提高算法的搜索质量。通过与3种常用启发式算法的对比实验,验证了HSNSGA-Ⅱ的有效性,并通过应用案例分析进一步验证了HSNSGA-Ⅱ的实用性。

关键词: 制造服务组合, 多目标优化, 服务能力受限, 非支配排序算法, 启发式搜索