GAILS: an effective multi-object job shop scheduler based on genetic algorithm and iterative local search

Area of Science:

• Operations Research
• Industrial Engineering
• Artificial Intelligence

Background:

• Job shop scheduling problems (JSSP) are crucial for smart factories but remain complex.
• Flexible JSSP (FJSSP) is more prevalent than classical JSSP in real-world scenarios.
• Existing methods often focus on classical JSSP, neglecting the nuances of FJSSP.

Purpose of the Study:

• To propose an effective hybrid method, GAILS, for solving both JSSP and FJSSP.
• To enhance resource management, production efficiency, and intelligent supply chain operations.
• To address the limitations of current research by focusing on more realistic FJSSP.

Main Methods:

• A hybrid approach integrating Genetic Algorithm (GA) for global solutions and Iterative Local Search (ILS) for local optimization.
• Encoding JSSP/FJSSP instances into machine and subtask sequences for GA.
• Utilizing multi-objective optimization (makespan, utilization ratio, maximum loading) to guide ILS.

Main Results:

• GAILS demonstrated strong search capacity, balancing global exploration and local exploitation.
• Comparative analysis on 66 public instances showed GAILS' effectiveness for both JSSP and FJSSP.
• The method achieved optimal-like solutions and outperformed state-of-the-art methods in most tested instances.

Conclusions:

• GAILS is a highly effective method for tackling complex JSSP and FJSSP.
• The proposed approach significantly improves key performance indicators such as makespan and resource utilization.
• GAILS offers a robust solution for enhancing smart factory operations and supply chain management.