Order acceptance model of multi-stage production based on simulated annealing algorithm

Often, company or factory has to refuse a number of profitable orders because of limited resources that cannot guarantee the delivery dates being ahead of the deadline of request. The lateness penalties of delivery always ruin the profitability of order, the extent to which is usually related to the length of lateness. In this paper, we extend the traditional single-stage model of order acceptance problem presented by Slotnick S. A.(1996) to multi-stage so as to make it be more close to the fact of order processing line, which is often staged. The Simulated Annealing algorithm is employed to find the optimal solution of the new model, and the detailed algorithm steps, as well as several feasible strategies on searching neighboring solutions, are given and tested together. Subsequently, we run the algorithm on simulating experiment data to validate its efficiency and effectiveness on finding optimal solution. The experiment results show that the multistage model can forecast the accepted orders being profitable or not more accurately than the traditional single-stage model. In the experiment, many orders that have been refused by traditional single-stage model is actually profitable and should be accepted in the simulation of multi-stage model. Copyright