Research On Supplier Selection Of A Company From The View Of Flexible Manufacturing

With the high degree of integration of informationization and industrialization,the data will replace energy as the raw material to create economic value,the industrial data will become a new impetus to industrial change,to promote the manufacturing model gradually moving from the traditional "push" production to order-driven "pull" production,so the flexible manufacturing has become inevitable development of enterprises.And how to quickly and adaptively respond to the procurement problem of order production are the keys to realize intelligent and flexible production.The change of the production mode will continue to promote the cooperation between the manufacturer and the supplier,and form a strategic alliance.The problem of the order assignment based on the purchase can be transformed into the process of selecting the supplier / supplier combination from the strategic alliance,we are collectively referred to as the order-driven supplier selection.Based on the practical problems of the NP company in the order-driven supplier selection,combined with existing research theories,methods and complexity of issues,proposed a new set programmer to meet the actual needs.Due to the complexity of supply and demand network and the ability of the agent to express different subjective goals,behavior and coordination,competition and cooperation well,this topic uses the multi-agent simulation technology to solve the problem of the supplier selection.Based on the existing based on quality,cost,service and other targets,and the supplier's comprehensive evaluation objective function is introduced to consider the supplier information and not increase the complexity of simulation calculation.Because the current simulation research is based on the multi-objective decision optimization,the multi-objective genetic algorithm(NSGA-II)is adopted to simulate the evolution of the Agent population,which can effectively solve the multi-objective decision problem.In addition,in order to reduce the subjectivity and industry limitation of the evaluation process,we introduce the random forest and the historical data,through the training and learning of the evaluation model,and obtains the optimal evaluation model and the key indicators and the indicator Weight of the optimal model.This can not only improve the rationality of supplier selection through historical data,but also provide support for supplier management.