A Linear Programming Approach for Performance Evaluation of Multi-Type Production Lines Applied in Manufacturing Strategies Comparison

Shengxuan Wang, University of Windsor


This research work develops an analytical approach to calculate the cycle time for each type of products in a multi-type serial production line with finite intermediate buffers and stochastic processing times. I consider the stochastic variables follow a certain distribution with mean and variance. The basic idea is to solve a linear programming approach modeling a production line operating with batch sized arrivals of different types of products and the cycle time can be found based on the batch restriction. A simulation model is created to test the relevance of the analytical approach and validate the proposed method's correctness. Scenarios such as switching processing from one product type to another without setup, with setup, and failure and repair are considered separately and comprehensive experiments combining these scenarios together are conducted as well. The failure and repair situation is stochastic as well. Experiment results are shown to validate the efficiency of the method. Periodic sampling approach is explored and considered while tackling these manufacturing strategies.