面向PID控制和仿真优化的集装箱码头作业调度

Container terminal task scheduling oriented to PID control and simulation optimization

鉴于现有运筹、仿真和优化方法应用于集装箱码头物流系统时均有一定的局限性,基于计算思维,将自动化科学中经典的微观PID控制机制和复杂系统研究中颇具前途的基于仿真的优化方法相结合,提出面向集装箱码头物流系统较为完备的作业调度计算体系和资源分配计算模式(其中调度策略和目标函数均基于PID控制思想进行定义),从而在允许的决策时间范围内获取动态不确定环境下港口计划调度的高质量多目标满意解。通过一个集装箱枢纽港的生产实例,对定义的PID调度算法(融合了处理器软亲和性策略)进行了参数整定计算实验,通过能力、任务延迟、资源利用和负载均衡等方面综合评估了参数优化后算法的性能,并给出了相应的结论。

Owing to the limitation that existing simulation and optimization method applied to Container Terminal Logistics Systems（CTLS）,based on computational thinking,the classical micro PID control mechanism and promising simulation-based optimization method in complex systems were combined to propose the complete task scheduling computational architecture and the relevant resource allocation computing mode oriented to CTLS,thus a high quality multi-objective satisfactory solution of planning and scheduling under dynamic and uncertain environments was obtained in the range of allowable decision time.Through a case of large-scale container terminal,the defined PID scheduling algorithm was designed and executed for parameters tuning,and the performance of algorithm by using optimal processing parameters was evaluated comprehensively from the perspective of traffic capacity,task latency,resource utilization and load balancing,and the corresponding conclusions were given.