A Novel Instance Generator for Benchmarking the Job Shop Scheduling Problem

The Job Shop Scheduling Problem (JSP) is a fundamental challenge in operations research, aiming to efficiently allocate tasks across machines to minimize criteria such as makespan, energy consumption, and tardiness. Benchmarking is crucial for evaluating algorithmic performance, with metrics like makespan and flowtime providing insights into scheduling efficiency. Instance characteristics, including job and machine numbers, processing time variability, and precedence relationships, influence problem complexity and solver performance. Evaluation of algorithm types, such as heuristics and metaheuristics, offers insights into their strengths and weaknesses. Reference libraries like JSPLIB enhance benchmarking efforts, while computational advancements drive the development of larger and more complex problem instances. Standardizing benchmarking procedures promotes consistency and reproducibility in algorithm evaluation. Our assessment of optimization solvers - OR-TOOLS, GUROBI, and CPLEX - underscores the critical role of instance generation in benchmarking. GUROBI consistently outperforms others, followed by CPLEX, while OR-TOOLS shows mixed results. Analysis of distribution types and release/due date settings provides valuable insights into solver behavior. Overall, rigorous experimentation and analysis are essential for advancing optimization methodologies and facilitating informed decision-making in problem-solving applications.