Deep Reinforcement Learning-Based Multi-Task Scheduling in Cloud Manufacturing under Different Task Arrival Modes

Abstract

Cloud manufacturing is a manufacturing model that aims to provide on-demand resources and services to consumers over the Internet. Scheduling is one of the core techniques for cloud manufacturing to achieve the aim. Multi-task scheduling with dynamical task arrivals is an important research issue in the area of cloud manufacturing scheduling. Many traditional algorithms such as the genetic algorithm (GA) and ant colony optimization algorithm (ACO) have been used to solve the issue, which, however, are either incapable of or perform poorly in tackling the problem. Deep reinforcement learning (DRL) that combines artificial neural networks with reinforcement learning provides an effective technique in this regard. In view of this, we employ a typical deep reinforcement learning algorithm – Deep Q-network (DQN) – and proposed a DQN-based multi-task scheduling approach for cloud manufacturing. Three different task arrival modes – arriving at the same time, arriving in random batches, and arriving one by one sequentially – are considered. Four baseline approaches including random scheduling, round-robin scheduling, earliest scheduling, and minimum execution time scheduling are investigated. A comparison of results indicates that the DQN-based scheduling approach is able to effectively address the multi-task scheduling problem in cloud manufacturing and performs best among all approaches.