论文部分内容阅读
传送带给料分批加工系统中,工件动态到达,并配置有存放待加工工件的缓冲区,其加工主体为批处理设备。考虑工件属性差异,重点研究单机模型的在线优化控制问题。以前视距离为控制变量,无穷时段内的工件处理率最大为优化目标,建立了系统的优化模型。针对该模型中的工件分批决策,提出一种以批处理机加工周期内加工能力浪费比最小为准则的工件分批规则。对于该模型中的行动选择决策,文中引入Q学习优化算法,以求解最优前视控制策略。通过仿真实验,对算法的有效性进行了验证,并分析了不同分批策略及参数对系统性能的影响。
Conveyor to the batch processing system, the workpiece arrived dynamically, and is equipped with a buffer to store the workpiece to be processed, the main processing for the batch processing equipment. Considering the difference of workpiece attributes, the paper focuses on the online optimization control problem of single machine model. Taking the distance of sight as the control variable and the maximum workpiece processing rate in an infinite period as the optimization objective, a system optimization model was established. Aiming at the batch work-piece decision-making in the model, a batch rule of work-piece based on the minimum waste-to-treat ratio in the machining cycle is proposed. For the action selection decision in this model, Q learning optimization algorithm is introduced in this paper to solve the optimal foreground control strategy. Through the simulation experiments, the validity of the algorithm is verified, and the impact of different batching strategies and parameters on system performance is analyzed.