Flow-Shop Scheduling Based on Reinforcement Learning Algorithm

Abstract

In the paper a machine learning based method will be proposed to give a quasioptimal solution to the /w-machine flow-shop scheduling problem. Namely, given a set of parts to be processed and a set of machines to carry out the process and the sequence of machines is fixed, each part should have the same technological path on all machines; the order of jobs can be arbitrary. The goal is to find appropriate sequence of jobs that minimizes the sum of machining idle times.