A Research About Glass Processing Warehouse Scheduling System Based On MAS

With the deepening of economic globalization,China has welcomed its own "Industry 4.0"--Made In China 2025.Once again,manufacturing has become a hot spot.The problem of industrial production scheduling is the core problem of industrial manufacturing,and it is also the academic hot spot studied by scholars all over the world.The efficiency of warehouse scheduling directly determines the quality of production scheduling.It is difficult to solve the warehouse scheduling problem under complex and uncertain task.In this paper,the multi-agent technology is combined with the actual situation of the production scheduling problem of glass processing manufacturing industry,and the warehouse scheduling system is designed effectively based on multi-agent technology.First,Our Research introduce the basic problem of warehouse scheduling and intelligent storage.Then the basic concept of Agent,the structure of Agent and the concept of multi-agent system are studied.This article expounds the advantages of Agent technology in scheduling problem.The glass processing and manufacturing warehouse system is mapped and modeled to the corresponding Agent.Then a multi-agent warehouse scheduling about glass system is established.Through designing evaluation index,the objective function is defined to realize the dynamic and global optimization of the system.Through the contract network negotiation strategy,make the system more intelligent.This article designs the corresponding algorithm strategy from the two industrial production aspects of processing which is equipment instruction execution and storage replenishment.The processing instruction is optimized by dynamic combination of different instructions.By introducing the concept of "information entropy" which comes from physics,the replenishment strategy can be foresighted based on the historical data.Through simulation experiment based on the real factory processing data simulated real scene modeling,the method of Agent results are compared with the general rule and the traditional dispatching rules,which prove Agent storage system has more feasibility and superiority.In the end,the multi-agent glass processing industrial original storage system,according to the evaluation index which dynamic split and combination of instructions.And by combining "information entropy" with scheduling selection,the predictive feeding algorithm of "space for time" makes the system better able to learn,adapt to itself,better performance and more flexible and robust.