针对作业车间调度问题中传统集中式调度方案的不足,提出一种在分布式环境下基于改进合同网协议的多智能体车间调度方法.在分布式制造系统环境下,将生产车间的设备进行Agent化,针对传统合同网协议通信负载大、效率较低,同时对分配方案没有评价机制,采用增强学习方式对任务分配过程进行强化学习,同时以制造车间的负载均衡为目标对合同网协议进行了相应改进.最后用仿真实验验证了方案的有效性.%Aiming at the flaws of traditional centralized means of shop scheduling problem,it proposes an improved Contract Net Protocol(ICNP) under the distributed system.As traditional CNP has low efficiency and no evaluation mechanism,it uses the reinforcement learning and designs an algorithm.The algorithm shows a concept about "distribution income" and is used for directed distribution with the assistant of the ICNP.Finally it takes a simulation experiment to test the performance of the algorithm,and verifies the feasibility and effectiveness.
展开▼