Job Shop Scheduling with Operators

（Learning Priority Rule Using Neural Network）

Abstract

This paper deals with a job shop scheduling problem that considers the skill difference of operators. A priority rule for this problem is constructed using a neural network. A neural network learns from input-output pairs extracted from a schedule optimized using an IP solver for small-scale problem instances. Through numerical experiments, it was found that trained neural network by the proposed method showed better performance than the existing priority rules even in larger scale problem instances.