Time bucket length and lot-splitting approach

Abstract

The effect of time bucket length on the choice of a lot-splitting approach is studied. Due to the continuing pressure to reduce throughput times and increase efficiency, managers apply various measures, such as lot splitting and cycle time reduction programmes, that change the length of the time bucket in their planning systems. It is shown that the choice for a suitable lot-splitting approach depends on the time bucket length used, so these measures cannot be treated independently as costs may increase rapidly without substantial effect on throughput times. Moreover, the current belief in the lot-splitting literature that the improvements of a variable lot-splitting approach are only marginal compared with an equal sub-batch approach is challenged. Some papers estimate the possible flow time improvement to be less than 5%. It is shown that such statements are not valid as the improvement depends strongly on the type of equal sub-batch strategy applied and the length of the time bucket. Moreover, flow time improvements are accompanied by extra transfer costs, so a cost perspective that includes both effects is more appropriate. The paper shows that equal sub-batching is only effective if information on the time bucket length and on product characteristics is included. At time bucket lengths between one and three days, costs are very sensitive for the type of sub-batching strategy applied. Simulation experiments show that variable sub-batching leads to average cost improvements from 3 to 11%, depending on the time bucket length and the equal sub-batch strategy applied. Lower costs are due to throughput time reductions and transfer cost savings.