Efficient local search algorithms for job-shop scheduling problems

Abstract

To solve job-shop scheduling problems, this paper develops four local search algorithms based on an input parameter of controlling the maximum allowable delay times of all machines. The use of this parameter in different values can change the size in numbers of the algorithm's solution space. The first proposed algorithm fixes the mentioned parameter's value and generates only forward schedules. The second algorithm reduces the parameter value iteratively and generates only forward schedules. The third algorithm reduces the parameter value iteratively; in addition, each operation-based permutation generated by this algorithm is transformed into both forward and backward schedules simultaneously. The fourth algorithm in the first-half iterations reduces the parameter value iteratively as well as generating only forward schedules; after that, this algorithm in the second-half iterations increases the parameter value iteratively as well as generating only backward schedules.