Research On Order Demand Prediction Based Dynamic Order Batch Scheduling Problem In The Cold Rolling Line

In view of low demand status of steel market,taking the cold rolling production process as background,an order demand prediction problem and an order batch scheduling problem with the consideration of forecasted order are investigated.The application of this research can achieve on-demand production and reduce the generation of inventory material.And it also can enlarge the optimization space of order batch scheduling,reduce production changeover cost,as well as transform the production line.For the order demand prediction problem,an order demand prediction model is formulated by using least square support vector machine.The differential evolution algorithm is adopted to optimize the parameters of SVM.Then,the order demand of the steel enterprise in the given time period can be forecasted.Based on the forecasted order demand,an integer programming model is formulated for the order batch scheduling problem in the cold rolling line with the consideration of order demand.An improved differential evolution algorithm is proposed to get the near-optimal solution in a short time.Finally,a support decision system is developed for the order batch scheduling process.The main works are as follows:1)In view of absence of production order causing by less order demand in the steel market,an order demand prediction problem is studied,providing a forward-looking guidance for the reasonable production scheduling of cold rolling.First of all,based on the actual production of cold rolling process,customer order is sorted according to their materials group.According to the analysis of historical order demand data,several order specification parameters,which have tight relationship with order demand,are selected.Based on the above order classification,the least square support vector machine was used to formulate the order demand prediction model under different material groups.In addition,the differential evolution algorithm was used to optimize the model parameters.The experimental results show that the proposed model is effective.Furthermore,the near-optimal sample size,which is of great reference value for practical data preprocessing,is obtained by analyzing the computational results.2)In view of the characteristics of batch production mode,high changeover cost and dynamic material arrival of cold rolling line,the demand prediction based dynamic order batch scheduling problem is studied.The task of the problem is to determine the production sequence of customer orders and forecasted orders in the given production period for each production unit of cold rolling line,considering practical constraints including the production capacity,inventory balance,and other techniques conditions.For this problem,with the objective of minimizing the production costs and load deviation cost,an integer programming model is formulated.The model for the small-sized instance can be directly solved by MIP solver CPLEX.Due to the fact that the CPLEX fails to solve the large instances of the problem,an improved differential evolution algorithm is developed to solve the problem effectively.By introducing the improved mutation and crossover strategy,the algorithm can avoid being trapped into premature convergence and local optimal.It can also enlarge the searching space and improve the search ability of the algorithm.The experimental results show that the proposed algorithm is superior to the basic differential evolution algorithm and the CPLEX in both optimization performance and solution speed.3)According to the actual demand of enterprises,the dynamic order batch scheduling decision support system for the cold rolling line is developed with the consideration of forecasted order demand.Through the good human-computer interaction interface,the system can forecast of order demand and generate order schedule for the cold rolling line.It can provide a good guidance for the production planning and scheduling in the cold rolling line,and hence improve the production efficiency.