Author:

Keywords:

Job shop scheduling, Decentralised scheduling, itec

Abstract:

We present two Reinforcement Learning approaches for the Parallel Machines Job Shop Scheduling Problem. The objective used is the minimization of the schedule makespan. We study two approaches, one where resources are modeled as intelligent agents and have to choose what operation to process next, and an other where operations themselves are seen as the agents that have to choose their mutual scheduling order. We use a value iteration method (QLearning) and a policy iteration method (Learning Automata). The results of both approaches improve on recently published results from the literature and we argue that they exhibit better scaling behaviour. We validate our approaches by applying them to the flexible job shop scheduling problem where operations can be executed on any of a number of available machines