Evaluation of solvers' performance for solving the flexible job-shop scheduling problem

• Main Author: Silva, João Pedro Martins Soares Castro
• Publication Date: 2023
• Other Authors: Lopes, Nuno, Alves, Cátia Filipa Veiga, Putnik, Goran D., Varela, M.L.R., Ferreira, Luis, Cruz-Cunha, Manuela
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: https://hdl.handle.net/1822/93372
• topic: Flexible Job-Shop Scheduling; Mathematical Optimization Solver

Scheduling is essential for the efficient planning of manufacturing enterprises and ultimately for its competitiveness. Cyber-physical production systems have been using heuristics to implement scheduling algorithms. Heuristics have the advantage of being fast to reach a solution, that is usually not the optimal one but close to optimal. Recent advances in mathematical optimization solvers have improved their performance and hence make them a contender to solve scheduling problems which were traditionally out of reach in terms of computation time feasibility. In this work, the authors compare the output and performance of two mathematical solvers, Z3 and Gurobi, with a game theoretic approach to solve a Flexible Job-Shop Scheduling Problem. To formulate the scheduling problem, an integer programming approach with the Picat programming language was used. The study found that the Z3 and Gurobi solvers outperformed the game theoretic approach, finding a better solution with a fast computation time for a relatively small but realistic problem size.