Variation in processing time in one-of-a-kind production: Case studies using the state space heuristic

Three criteria are important to evaluate a heuristic for adaptive production control, which are optimality, computational complexity and flexibility. However, the stability of a schedule generated by a heuristic is also important from the perspective of manufacturing systems. In this paper, we carry out case studies on variation in processing time to examine the stability of the state space (SS) heuristic proposed by Li et al., 2011, and verify some former conclusions by Cao et al., 2005 about the effect of variation in processing time on production. Our results show that (1) our SS heuristic is stable to the disturbance of variation in processing time; (2) the estimation error does have a trivial effect on production, which confirms Cao et al.'s conclusion; (3) the variation of actual processing times might not have more important effect than the estimation error if we use a suitable heuristic, which is contrary to Cao et al.'s conclusion. More importantly, our results show that enlarging processing times of some jobs on some stages may result in a smaller makespan.