Loss function based robust scaling parameters for composite dispatching rule ATCS

Abstract

Composite dispatching rules are of great importance in solving realistic scheduling problems because they are fast and easy to implement. Under some circumstances, a quick and easy solution is needed for prompt implementation, in which the dynamic scaling parameter value adjustment, which is required by the fitted function methods in the literature, may not be allowed. On the other hand, for a given scheduling system, the problem instances that appear over time may have a lot in common, such as similar setup times or due date assignments. If the most frequent region of the problem instance is known in advance, the appropriate robust scaling parameters can be found using loss function for good and stable performance. In this paper, a loss function based method is proposed to find the robust scaling parameter values for Apparent Tardiness Cost with Setups (ATCS) rule under such circumstances. The computation results show that the identified robust scaling parameters (k1= 0.8, k2 = 0.1) achieve better performance than the fitted function method in the literature.