首页|期刊导航|模具技术|基于改进GA与DDQN的模具加工智能调度及系统设计

基于改进GA与DDQN的模具加工智能调度及系统设计OA

Intelligent scheduling and system design for mold processing based on improved GA and DDQN

中文摘要英文摘要

为提高模具加工的生产效率和稳定性,提出一种基于改进遗传算法(GA)与双重深度 Q 网络(DDQN)的模具加工智能调度及系统设计.首先考虑模具加工设备静态不变条件,以模具加工成本最小为目标函数,采用初始种群和交叉变异改进的GA 进行求解,实现了模具加工静态调度优化;然后考虑模具加工设备动态调整条件,以最小化模具加工径流和为目标函数,采用 DDQN 进行求解,实现了模具加工动态调度优化.最后,基于 Visual Studio 软件和 C#语言构建模具加工智能调度系统,对方法进行验证.结果表明,改进 GA 对模具加工静态调度的总成本为 874.6 元,平均相对偏差率为 0.01%;DDQN对模具加工动态调度的平均时间为 7.48 s;基于改进GA与 DDQN 的模具加工智能调度系统,可提高模具加工生产效率和稳定性,减少模具加工时间.由此得出,本方法及系统具有一定实际应用价值,为模具加工的智能调度提供了参考.

To improve the production efficiency and stability of mold processing,an intelligent scheduling and system design for mold processing based on improved genetic algorithm(GA)and double deep Q-network(DDQN)is proposed.Firstly,considering the static invariance condition of the mold processing equipment,with the objective function of minimizing the mold processing cost,the initial population and cross mutation improved GA were used to solve the problem,achieving the optimization of static scheduling for mold processing;Then,considering the dynamic adjustment conditions of the mold processing equipment,with minimizing the mold processing runoff as the objective function,the DDQN was used for solving,achieving dynamic scheduling optimization of mold processing.Finally,a mold processing intelligent scheduling system was constructed based on Visual Studio software and C#language to validate the method.The results show that the total cost of improving the GA for static scheduling of mold processing is 874.6 yuan,with an average relative deviation rate of 0.01%;The average time for dynamic scheduling of mold processing by DDQN is 7.48 s.Based on improved GA and DDQN,an intelligent scheduling system for mold processing can improve production efficiency and stability,and reduce mold processing time.From this,it can be concluded that this method and system have certain practical application value,providing a reference for intelligent scheduling of mold processing.

郑素萍;许严翀;卢明

巨鑫建设集团有限公司余杭分公司,浙江 杭州 311100纽约大学,美国 纽约 11101巨鑫建设集团有限公司余杭分公司,浙江 杭州 311100

信息技术与安全科学

模具加工智能调度静态调度动态调度遗传算法(GA)双深度Q网络(DDQN)

mold processingintelligent schedulingstatic schedulingdynamic schedulinggenetic algorithm(GA)algorithmdouble deep Q-network(DDQN)network

《模具技术》 2026 (4)

36-44,9

评论