Abstract
In the production environment, the goal of a successful scheduling system is to reduce inventory cost and shortage cost. However, it is difficult to make lot size decision under a time-varying demand. Facing a capacitated lot size scheduling problem of color filters manufacturing, this study tries to establish an effective method to reduce inventory and avoid backlog. We decompose a multi-stage problem into several single stage sub-problems, and we use the input requirements of a later stage as the demands of its previous stage. To solve each sub-problem this study proposes a heuristic approach. To avoid over-estimating capacity, this study divides production scheduling horizon into many small segmented buckets. To reflect the time-varying property of demands, this study decides initial production order by using “run out time of the inventory” proposed by Leachman【1988】. Then, we use “exchange” and “merge” methods to adjust production sequence, and rationalize the solutions. During the above procedure, this study calculates an accurate total cost, ensures that the calculated lot size decision can reduce inventory, shortage and setup cost. The above procedure will be iterated until a satisfied reasonable solution is obtained. Experimental results indicate that such a methodology can effectively obtain a production schedule for each type of product to reduce total inventory, backorder and setup costs if choose a appropriate number of segmented buckets. Longer production buckets will lead to best solution, but it is time consuming. Shorter segmented buckets will decrease time for solving problem, and get a satisfied solution. According to the experimental results, we suggest the length of the segmented buckets is about 2-3 times average setup time.