离散型生产排序的自适应控制

Adaptive Control of Discrete Production Scheduling

离散型生产活动排序的自适应控制系统包含预测器、最优控制器、参数估计器三部分。基于极大极小代数法的预测器计算各种排序方案的完工时间,该算法以任务投入加工的时间作为状态变量,以预测的任务加工的完工时间作为输出变量,通过输入工艺路线矩阵、加工顺序矩阵和加工时间矩阵,根据生产活动状态变迁的规则建立状态变量的递推计算公式,经过多重循环求出所有状态变量的取值;最优控制器以加工顺序矩阵为控制变量,利用预测器逐步试探求解,最后稳定在性能指标最优的加工顺序矩阵的方案上;参数估计器对加工工时定额进行自适应地学习,利用反馈的真实的加工工时,不断修正定额并收敛到平均值。文中阐述了这种控制方法的仿真实例。

The adaptive control system of discrete production scheduling includes forecast ware, optimum controller and parameter estimate ware. Based on the max＆min algebra algorithm, the forecast ware computes the end time of tasks of various schemes and this algorithm regards beginning time for processing parts as state variable with the end time forecasted of processing parts as variable exported. Through routing matrix, processing order matrix and the matrix of handling time, i according to the change rules of production campaign state, the recurrence formula is established, by which all state variables are computed. Optimum controller takes processing order matrix as control variable, and sounds out solves step by step through the forecast ware and stabilizes the scheme in processing order matrix with optimum performance finally. Parameter estimated ware emulates the actual handling time of feedback adaptively, convergences to average of handling time quota unceasingly. A simulation example of this kind of cona＇ol method was elaborated.