基于多智能体仿真的高速工具钢炼钢计划调度  

High-Speed Tool Steel Steelmaking Scheduling Based On Multi-Agent Simulation

作　　者：梁青艳[1] 孙彦广[1] LIANG Qing-yan;SUN Yan-guang(Automation Research and Design Institute of Metallurgical Industry,China Iron and Steel Research Institute Group,Beijing 100071,China)

机构地区：[1]中国钢研科技集团冶金自动化研究设计院,北京100071

出　　处：《计算机仿真》2021年第6期367-372,共6页Computer Simulation

基　　金：国家重点研发计划专项资助(2017YFB0304000)。

摘　　要：为了有效解决具有强耦合、多目标、多约束、多干扰特性的高速工具钢生产流程中计划调度较为困难的问题,提出一种基于多智能体技术的计划调度建模方法。把钢铁生产流程抽象为一个多智能体系统,对复杂物流系统进行建模与仿真优化,通过多目标评判与迭代优化获取作业计划的满意解,计划下达后通过对执行偏差的反馈,进行动态调整,提高计划的适应性和执行性。最后通过仿真对不同的工况场景进行模拟,均能给出可行解决方案。实验表明所建模型能大大减少无效作业时间,提高作业效率,并且能适应多变的现场环境。In order to effectively solve the problem of planning and scheduling in the high-speed tool steel production process with strong coupling, multi-objective, multi-constraint, and multi-interference characteristics, a method of modeling and simulation optimization of complex logistics system based on multi-agent technology was proposed. The steel production process was abstracted into a multi-agent system. Through the multi-objective evaluation and iterative optimization, a satisfactory solution to the operation plan was obtained. After the plan was released, the feedback of the execution deviation was dynamically adjusted to improve the adaptability and execution of the plan. Finally, simulations were used to simulate different working conditions, which could give feasible solutions. Experiments show that the built model can greatly reduce invalid operation time, improve work efficiency, and adapt to the changing field environment.

关 键 词：生产计划 多智能体 多目标优化 动态调度 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]