热轧轧制计划的多目标优化模型及算法  

Multi-objective optimization model and algorithm for the hot rolling batch scheduling problem

作　　者：贾树晋[1] 李维刚[1] 杜斌[1] 

机构地区：[1]宝钢集团中央研究院自动化研究所,上海201900

出　　处：《武汉科技大学学报》2015年第1期16-22,共7页Journal of Wuhan University of Science and Technology

摘　　要：针对热轧轧制计划优化问题,建立基于奖金收集车辆路径问题(PCVRP)的多目标优化模型,其中包含两个目标:目标1为最小化相邻板坯的宽度、厚度与硬度的跳跃惩罚;目标2为最大化收集的奖金,即使得尽可能多的板坯编入轧制计划。在此基础上,提出一种基于Pareto最优的多目标蚁群系统算法(MOACS),避免了传统加权法需要确定目标权重系数的缺点,一次运行可产生多个Pareto最优解,给决策者带来了更大的决策自由度。现场数据测试表明该算法具有良好的优化性能和实用性。In view of the hot rolling batch scheduling problem,a multi-objective prize collecting vehicle routing problem（PCVRP）model is presented.This model has two objectives：the first objective is to minimize the penalties caused by jumps in width,gauge and hardness between adjacent slabs,and the second one is to maximize the collected prize,which means more slabs can be inserted into the rolling batch.Then,a multi-objective ant colony system algorithm based on Pareto optimal is proposed to solve the PCVRP model.This algorithm can not only avoid the selection of weight coefficients encountered in the single objective optimization but also obtain multiple Pareto-optimal solution,which provides more decision-making flexibility for the schedulers.Large numbers of site tests show that this algorithm has optimal performance and good practicability.

关 键 词：热轧轧制计划 多目标优化 PARETO最优 蚁群算法 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]