A Genetic Algorithm for tough Scheduling Problems

摘要

A new selection method for Genetic Algorithm used to solve a tough combinatorial optimization problem has recently being proposed. The method consists of a partial enumeration of neighbors of a set of sampled individuals to control the algorithm selection pressure. Here, we improve this selection method by including problem specific knowledge in the neighborhood generation procedure. The problem considered is the Job Shop Scheduling Problem. Particularly, a set of still open benchmark problems are dealt with. The attained accuracy results are comparable with standard existing methods available in the literature.