This project was developed by me under the orientation of Prof. Dr. Marcone Jamilson Freitas Souza in Federal University of Ouro Preto, Brazil and tried to solve the Job Shop Scheduing Problem with total weighted tardiness.
The considered problem representation was based on a disjuctive graph model which each job operation is equivalent a vertex and a restriction between two operations generates an edge; Two dummy nodes were added to help the solution evaluation. Originally this representation was extracted from (Kuhpfahl and Bierwirth, 2016).
To create a initial solution was tested many dispatch rules and those that presented the best results were: ASPRT, MOD, SPRT and WSPT rules.
To explore the search space was implemented some algorithms: "Variable Neighborhood Search", "Smart Variable Neighborhood Search", GRASP, and "Simulated Annealing". Also, some local searches were implemented: "Best Improvent" and "First Improvement".
All methods mentioned above are available in this repository.
Any question you cand find me on: [email protected]