Production Planning of a Multi-Item-Single-Facility System with Non-Stationary Demands

We consider a single production facility capable of producing N products each demanded at a varying rate throughout the planning horizon. The production facility makes a fixed lot in a given period. No more than one product can be produced in a period. Given the initial inventory for each product and a finite planning horizon, the objective is to identify the production schedule with the minimum total inventory holding and backordering costs. Dominance relations between the products are derived, and strong dynamic lower bounds on the optimal cost are developed. Together, they are used in an efficient dynamic branch and bound algorithm to generate the optimal schedule. The computational results show that the proposed branch and bound is effective in reducing the CPU time, hence allowing for the solution of large size problems which is not possible by using other methods