考虑学习效应和能源消耗的并行机调度研究  

Multi-Objective Parallel Machine Scheduling Considering Learning Effect and Energy Consumption

作　　者：耿瑗璐 杨玉中[1] GENG Yuan-lu;YANG Yu-zhong(School of Energy Science&Engineering,He’nan Polytechnic University,He’nan Jiaozuo 454000,China)

机构地区：[1]河南理工大学能源科学与工程学院,河南焦作454000

出　　处：《机械设计与制造》2024年第9期130-135,共6页Machinery Design & Manufacture

基　　金：国家自然科学基金项目(51874121)。

摘　　要：学习效应广泛存在于车间生产,基于学习效应的车间调度优化准确度更高。运用理论分析和遗传算法对并行机调度问题进行了研究。构建了以最大完工时间和能耗最小为优化目标的多目标并行机调度模型,设计了一种基于贪婪解码算法的第二代非支配排序遗传算法。基于工序和机器进行双层实数编码,解码过程插入贪婪算法,基于工件工序号,得到该道工序在当前设备中待加工序号及当前设备存在的加工间隙,在设备间隙矩阵中判断工件工序可插入的位置,进而更新开始及完工时间与机器能耗使用情况。通过数值算例验证了模型及算法的有效性。Learning effect exists widely in workshop production,and the accuracy of shop scheduling optimization based on learn⁃ing effect is higher.The parallel machine scheduling problem is studied by theoretical analysis and genetic algorithm.A schedul⁃ing model of multi-objective parallel machines with maximal completion time and minimum energy consumption is established.A second-generation non-dominated sorting genetic algorithm based on greedy decoding algorithm is designed.Based on process and machine double real number encoding,decoding process to insert the greedy algorithm,based on the workpiece process,get the procedure in the current equipment for processing the serial number and the machining gap existing in the current equipment,the equipment matrix to judge the position of the workpiece process can be inserted into the gap,and then update the start and fin⁃ish time and the energy consumption of machine usage.A numerical example is given to verify the effectiveness of the model and algorithm.

关 键 词：并行机调度 NSGA-Ⅱ 贪婪算法 学习效应 最小能耗 

分 类 号：TH16[机械工程—机械制造及自动化] TP301[自动化与计算机技术—计算机系统结构]