Research On Production Scheduling Problem Of Cold Rolling Line Based On Multi-objective Differential Evolution Algorithm

The preparation of the townhouse plan for the production of cold rolling steel coils is to determine the materials to be produced on each crew and ascertain the order of material production according to the inventory of each process,it's an important part of ensuring the rationality of steel production.It involves multiple processes,such as units and materials,which is a large-scale NP-hard problem.In the process of continuous production,the frequent varieties switching of the connecting materials will incur additional costs,and even affect the quality of the products,then reduce customer satisfaction.In a production line,the large production sequence difference of materials of different brand but same kind producing in connecting process will affect the coherence of the unit production.At the same time,to ensure the rational use of enterprise resources,the capacity of the unit should be maximized.However,in the current production plan of the enterprise,it is difficult to fully consider these factors.Therefore,the research on the cold rolling production line joint planning problem plays an important role in the improvement of product quality and production efficiency.The innovation of this thesis mainly includes two points.In this thesis,the production line of a Silicon steel cold rolling production line of a large steel mill was considered.First of all,unlike the previous research based on single process,this problem studies the scheduling of two connecting processes which are acid rolling unit and continuous retreating unit.And considering the relationship between multiple targets,the descriptive mathematical model based on multi-objective with material selection is established.Secondly,for the difficulties in solving multi-objective mathematical models,an improved differential evolution algorithm based on three strategies is designed.The main research is as follows:1)Based on the principles and process constraints of the cold rolling scheduling of production line,the problem of the scheduling planning with material selection is extracted combined with the method and problems of manual preparation plan.Considering the commonalities of the two processes in the planning process,the theoretical structure of the scheduling problem is provided and a model of multi-objective mathematical programming is established.The objection includes there parts.The first is to minimize the switching cost.The second is to reduce the production order difference between the front process and back process.The last one is to maximize unit capacity.Constraints such as unit production capacity,process specifications,and uniqueness are considered in the model.2)The problem of this thesis is solved using the improved differential evolution algorithm MKDE.Three improved strategy optimization algorithms are proposed based on the characteristics of the problem:the coefficient of variation used the element-based parameter strategy;the search direction which guides the orientation-based mutation strategy;the individual selection based on K proximity.Compared with DE algorithm,NSGA-? and MOGLS algorithm based on the three metrics(convergence,dominance,and dispersion),numerical experiments show that the algorithm is effective.3)Combined with scheduling requirements of the enterprise,scheduling planning of the cold rolling production line of decision support system was designed and developed in detail.The mathematical model and the improved differential evolution algorithm proposed in this thesis was embedded to realize data management,parameter setting,planning,result evaluation,etc.Through the interface of the system,human-computer interaction can be carried out simply and conveniently,and the plan can be automatically prepared to flexibly guide the production work on site,and the production cost of the enterprise can be effectively reduced,and the production efficiency can be improved.