Research On Dynamic Matching Problem Between Coils And Orders

In this thesis,the dynamic matching problem between coil and order is studied in the background of coil production process.This thesis discusses the dynamic matching problem between the coil and the order from the perspective of integration of the customer-order coil and the open-order coil,which extends the optimization space from the two dimensions of matching objects and time,improves the profit as well as improves customer satisfaction and competition of iron and steel enterprises.Aiming at the problem of dynamic matching between coil and order,an integer programming model is established,and an improved differential evolution algorithm is proposed to quickly obtain the near optimal solution.The proposed hybrid mutation strategy and algorithm parameter adaptive strategy can significantly accelerate the convergence rate of the algorithm.In order to adapt to the complex and varied working conditions of steel enterprises,a rolling time-domain mechanism is designed to reduce the impact of dynamic disturbance factors on the matching scheme.The main contents of this paper are as follows:(1)The dynamic matching problem between coils and orders is derived from the coil production process in the iron and steel enterprises.The task of this problem is to optimally allocate both the non-delegated coils and assigned coils to the orders within the given time horizon.During the decision making,the constraints of the capacity of an order's demand,inventory balance,and availability of coils should be considered.Through the integrated optimization of non-delegated matching and full-volume matching,the optimization space can be much extended.Through the analysis of practical production process,an integer programming model that reflects the practical production is formulated for the integrated optimization problem,in which the maximization of coil utilization and the minimization of production and inventory are taken as the objective.Finally,the validity of the model is verified by numerical experiments.(2)A differential evolution algorithm is proposed to efficiently solve the dynamic matching problem between actual coils and orders,which can handle the situation that standard commercial software cannot solve large scale problems.In order to prevent the algorithm from being trapped in local optimality,a hybrid mutation method with DE/rand/1 and DE/best/1 and a modified adaptive strategy are proposed.An individual improvement strategy is also proposed to improve the searching ability of the algorithm.Finally,the effectiveness of the algorithm and the improved strategies is verified by numerical experiments based on different scales of problems and comparison with CPLEX.(3)In order to handle the frequent fluctuations of production conditions,the dynamic matching problem between coils and orders is embedded into the framework of rolling horizon optimization so as to reduce the impact of unpredictable dynamic disturbances on the matching decisions.Through incorporating the improved differential evolution algorithm into the rolling horizon framework,a rolling horizon optimization mechanism reflecting practical production conditions is proposed.At last,the impact of the number of rolling windows on the performance of global rolling horizon optimization is analyzed by different numerical experiments.(4)Based on the above mathematical models and algorithm,a subsystem of the coil and order matching system is developed.