论文部分内容阅读
研究一种两类品种工件混流的多站点传送带给料加工站系统的优化控制问题.系统中的站点如何协同工作完成工件加工任务,是提高系统生产率的重要课题.将前视距离作为各站点的决策变量,通过站点间的局部信息交互,提出一种品种均衡工作模式,并运用一种模型无关的串行反馈式多agent强化学习算法求解系统的最优策略.实验结果验证了该工作模式的合理性和算法的有效性,并分析了部分参数变化对系统性能的影响.
This paper studies the optimization control of a multi-site conveyor feeding system with two types of workpieces, and how to coordinate the work of site processing tasks in the system is an important issue to improve system productivity. Decision-making variables, through the site of local information exchange, a variety of balanced working model, and use a model-independent serial feedback multi-agent reinforcement learning algorithm to solve the system’s optimal strategy.The experimental results verify the work mode Rationality and the effectiveness of the algorithm, and analyzes the impact of some parameters on system performance.