| This dissertation, with financial support by science and technology special project of Zhejiang province, studys the workshop scheduling problems in the environment of MES and multi-agent system based on the analysis of discrete production characteristics and requirements. The main research works are as follows:1. Establishes the MES scheduling framework model based on MES framework model. Designs the MES scheduling process based on the prediction and reactive scheduling strategy and the structruction of MES scheduling framework model. In addition, establishes models of data acquisition and agent scheduling.2. Gives the dynamic time and handing time determining method, structs the mixed workshop static scheduling model under the environment of MES and the analysis of discrete enterprise production characteristics. And puts forward the improved immune genetic algorithm mixed with the strategy of decomposition.3. Gives the dynamic rescheduling assumption conditions and targets, analyzes dynamic rescheduling diffusion effect. Designs a dynamic rescheduling method based on improved rolling window technology, then proposes one kind of adaptive evolutionary programming algorithm based on improved Q-learning (QEP), and mixes with the Contract Net Protocal mechanism (QEP-CNP). Simultaneously, gives the artificial rescheduling method.4. Completes the system development basing on ASP.NET and B/S structure, combines absorber enterprise example to show the result of the system running, finally, realizes absorber production scheduling optimization. |
