Research On Fuzzy Hierarchical Production Planning And Scheduling Methods In Multi-product Enterprise With Uncertain Demand

With the development of the economy and diversification of consumer demand, the production mode of large quantities and less variety in the past are replaced by high mix. How to allocate the production quantity in each period and develop production plans accurately is the urgent problem most of enterprises need to face. Hierarchical Production Planning is a well-known approach to cope with production planning and scheduling problems in multi-product enterprise. Due to the uncertainty of market demand, the determined HPP simply cannot provide a feasible and efficient production plan for the enterprise. The thesis focuses on the research on fuzzy hierarchical production planning and scheduling methods with uncertain demand.Firstly, considered the characteristics of discrete manufacturers and uncertain demand, we incorporate the fuzzy variable into the HPP structure to handle the uncertainty and infeasibility issues. A fuzzy hierarchical production planning and scheduling framework is proposed which divides production planning and scheduling problems into four layer, namely fuzzy aggregate production planning(FAPP) layer, fuzzy family disaggregate production planning(FFDP) layer, production operations sequence(POS) layer and job shop scheduling(JSS) layer.Secondly, the research establishes corresponding mathematical models for the first three layers respectively. Fuzzy demand, inventory and subcontract factors are taken into account in the FAPP layer. A mixed-integer linear programming model is built in this to determine the production, inventory, backorder and subcontract of each product-family. The FFDP decomposes the FAPP. A fuzzy weighted integer goal programming model is designed to solve the production volume and optimal production sort of each product family in a short period. And thus the FAPP model would be detected and adjusted; At the POS layer, a 0-1 integer programming model is established to obtain the production operation sequence planning.Then, focused on the job shop scheduling layer, the constraint satisfaction technique is used to model the job shop scheduling problem. A hybrid algorithm which combines constraint propagation and branch and bound is designed to solve the model. As a result, the job sequence for the parts of product is obtained.At last, the production situation of a LED lighting manufacturer is introduced. We build mathematical models correspondingly on the basis of the product demand and capacity of the enterprise. And then use Xpress-MP which is the operations research optimization software to solve them. The results show that the proposed methods on production planning and scheduling provide the enterprise with a production decision efficiently.