A block approach to earliness-tardiness scheduling problems

In this paper, we consider a single-machine job scheduling problem where the objective is to minimize the weighted sum of earliness and tardiness (E/T) penalties of jobs. This problem is consistent with the just-in-time (JIT) production. We propose partitioning of permutation into subsequences (blocks) and replacing sets of moves with its representatives, significantly decreasing the size of the searched neighborhood. Some new properties of the problem and compound moves make eliminating "bad" elements and speeding up calculations possible. These properties allow us to propose a new fast local search algorithm based on a tabu search method. Computational experiments are presented and the results show that the algorithm proposed allows us to obtain the best-known results in a short time.