Analysis of rolling horizon method and end-of-horizon effects in aggregate planning systems

Abstract

A computer model is built to simulate master production scheduling activities in uncapacitated single item single level production system under demand uncertainty in a rolling time horizon environment. The output from the simulation is examined to decide whether assigning a value for ending inventory would make a change in the relative effectiveness and performance of methods. To assess improvements in total cost when a lower bound is assigned for the ending inventory level, the level of ending inventory is set at three levels. Moreover, the effect of MPS system design parameters on the effectiveness of proposed solution method is investigated. Effectiveness is measured by making comparison with a benchmark policy that allows zero ending inventory as usual method. The results of the study show that extending the horizon and assigning a lower bound for the ending inventory have impacts on the performance measure of cost. Especially when the demand variability is low, cost structure is minimum, freezing proportions of 0.25 and 0.75 is used and plan length is small, proposed solution method appears to provide better performance in terms of cost than the usual method of zero ending inventory. Furthermore, the selection of the master production schedule freezing parameters is also significantly influenced by solution type used. The findings from this study can help planners optimize their production plans by selecting more reasonable and easy to implement methods and scheduling parameters, thus improve the performance of production systems.