Abstract
The goal of a Just-in-time production system is to minimize the inventory cost. However, it is difficulty to make lot size decisions under a dynamic demand environment, whose demands vary with time. To achieve the goal of JIT, this study considers multi-machine dynamic-demand capacitated lot size scheduling problem with the objective of minimizing the inventory cost. We divide the overall production decision into two kinds of decisions: “discrete decisions” and “continuous decisions”. Discrete decisions include the product type for each of the production runs, the beginning period of each run, the ending period of each run, and product on the machines. The continuous decisions suggest the precise beginning epoch and ending epoch of each run. We will utilize tabu search and simulated annealing to determine the discrete decisions. Then, by using the discrete decisions, a mathematical programming model, proposed herein, will be used to solve the continuous decisions. The above procedure will be iterated to solve our problem.Experimental results indicate that such a methodology can effectively obtain a production schedule for each type of product to reduce total inventory costs and eliminate backorders. Experimental designs and statistical methods are used to evaluate and analyze the performance of tabu search and simulated annealing. As a result, tabu search performs significantly better than simulated annealing. Therefore, we suggest that tabu search is utilized to search for discrete decisions.