Open Access

ARTICLE

An Advantage Actor-Critic Approach for Energy-Conscious Scheduling in Flexible Job Shops

Saurabh Sanjay Singh^*, Rahul Joshi, Deepak Gupta

Department of Industrial, Systems, and Manufacturing Engineering, College of Engineering, Wichita State University, Wichita, KS 67260, USA

* Corresponding Author: Saurabh Sanjay Singh. Email:

Journal on Artificial Intelligence 2025, 7, 177-203. https://doi.org/10.32604/jai.2025.065078

Received 03 March 2025; Accepted 04 June 2025; Issue published 30 June 2025

Abstract

This paper addresses the challenge of energy-conscious scheduling in modern manufacturing by formulating and solving the Energy-Conscious Flexible Job Shop Scheduling Problem. In this problem, each job has a fixed sequence of operations to be performed on parallel machines, and each operation can be assigned to any capable machine. The problem statement aims to schedule every job in a way that minimizes the total energy consumption of the job shop. The paper’s primary objective is to develop a reinforcement learning-based scheduling framework using the Advantage Actor-Critic algorithm to generate energy-efficient schedules that are computationally fast and feasible across diverse job shop scenarios and instance sizes. The scheduling framework captures detailed energy consumption factors, including processing, setup, transportation, idle periods, and machine turn-on events. Machines are modeled with multiple slots to enable parallel operations, and the environment accounts for energy-related dynamics such as machine shutdowns after extended idle time, limited shutdown frequency, and machine-state transitions through heat-up and cool-down phases. Experiments were conducted on 20 benchmark instances extended with three energy-conscious penalty levels: the control level, moderate treatment level, and extreme condition. Results show that the proposed approach consistently produces feasible schedules across all tested benchmark instances. Relative to a MILP baseline, it achieves 30%–80% lower energy consumption on larger instances, maintains 100% feasibility (vs. MILP’s 75%), and solves each instance in under 0.47 s. This work contributes to sustainable and intelligent manufacturing practices, supporting the objectives of Industry 4.0.

---

Keywords

---