A multi-stage stochastic linear programming model for apparel production planning

Abstract

This research is aimed at developing a multi-stage stochastic linear programming model in apparel production planning. Multiple constraints on the capacity of labor, machine and quota are all considered in the production planning process in multiple periods and also in multiple stages in stochastic linear programming model. The objective of stochastic linear programming model for apparel aggregate production planning is to minimize the manufacturing cost such as in-house manufacturing and subcontracting, in addition, to reduce pre-build inventory, in which supply and demand will meet under constraints in multiple periods. Due to the short supply, demands cannot be satisfied which result in penalty of order cancellation. Over supply may also create a burden of inventory. Holding inventory certainly makes an impact on final cost. The stochastic linear programming model in multiple stages and multiple levels can calculate the number of in-house manufacturing and subcontracting in each production period. The stochastic linear programming model should be implemented since the formulae in iLOG CPLEX Studio programming is designed for importing production schedule date and criteria of constraints. Afterwards, we can make use of the data to minimize the total costs of production plan and thus to optimize the manufacturing resources (labor and machine) and quota logistically. At the application level, the result of the production plan after implementation of multi-stage stochastic linear programming model in iLOG CPLEX Studio is satisfactory. It shows a cost reduction in comparison with the cost of current production plan. Thus it proves the success in minimizing the cost of manufacturing. In this way, it clearly states that the problem of production planning in apparel factories can be resolved by planning manufacturing quantity in each factory and subcontracting based on criteria of constraints in multiple periods. Given this modest success, the methodology of this project attains its goal as it meets customers’ expectation and satisfies customers’ demands with on-time shipment deliveries based on good practice production planning and management. An optimal solution in terms of multi-stage of stochastic linear programming model for apparel production planning is found to resolve problems under constraints successfully. The goal of the project is achieved with proven cost saving result.

Date of Award	2 Oct 2008
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Chuangyin DANG (Supervisor)