多目标下航天产品生产车间柔性资源配置与调度集成优化

Multi-objective integrated optimization of flexible resource allocation and scheduling in the aerospace production workshop

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Original languageChinese (Simplified)
Journal / Publication中国管理科学
Publication statusOnline published - 26 Apr 2022

Abstract

航天产品的生产具有多品种、小批量特点,根据生产计划进行柔性资源配置与调度,对生产效率和响应能力都有重要影响。考虑了生产资源投入量可变、生产中存在准备加工运输等多种时间、生产资源性能差异等情况,本文以最小化超期交付天数、生产成本和总投入资源数量为目标进行了柔性资源配置与调度集成优化问题的研究。首先,本文对研究问题进行了界定和详细描述,建立了多目标规划模型。其次,鉴于该问题的 NP-hard 特点设计了基于 Pareto 优化的改进多目标差分进化算法,通过启发式染色体生成策略得到高质量的初始解集,结合快速非支配排序方法对个体进行排序,利用差分策略进行迭代优化,最终获得满意的解集。最后,通过算例分析将本文算法与两种常用算法进行了对比分析,验证了算法的有效性,并得出:当任务量多生产资源不足时,增加资源数量可以减少超期时间和生产成本;但当任务量少生产资源充足时,即使投入更多资源,也不会使得超期时间和生产成本进一步减少,还可能会带来资源浪费。
To cater to the multi-variety and small-batch characteristics of aerospace product manufacturing, flexible allocation and scheduling of resources in production workshops based on producing planning is key to maintaining efficiency and response capability. In particular, this paper formulates and solves an integrated optimization problem of resource allocation and scheduling to minimize production delay, manufacturing costs, and total resources across multiple scenarios with various production resources, in-workshop transportation, and resource availability. A multi-objective programming model is first formulated based on a detailed delineation of our research questions. We then put forward an improved multi-objective differential evolution (IMODE) algorithm based on Pareto optimization to solve such an NP-hard problem efficiently. Upon obtaining a high-quality set of initial chromosomes based on heuristic search, the final solution of a Pareto set is derived from fast non-dominated sorting followed by iterative optimization through a differential evolution algorithm. Based on 11 scenarios with different combinations of job numbers and due dates generated from empirical data, IMODE on average outperforms NSGA-II and MOPSO, two widely-adopted algorithms for multi-objective optimization. Furthermore, our additional analyses identify a ceiling effect of resource quantity on overdue days and manufacturing costs. When there is a large number of jobs with insufficient resources, on the one hand, increasing the number of resources can mitigate production delays and costs. On the other hand, the allocation of more resources is unable to shorten overdue days or bring down costs in the context of a small job number with adequate production resources. In sum, flexible integrated resource allocation and scheduling through IMODE enhance resource utilization and manufacturing efficiency in the face of dynamic production demands.

Research Area(s)

  • 生产车间, 柔性资源配置与调度, 多目标优化, 自适应差分进化算法, 快速非支配排序, 启发式策略, production workshop, flexible resource allocation and scheduling, multi-objective optimization, adaptive differential evolution algorithm, fast non-dominated sorting, heuristic strategy