Resumen
A simple genetic algorithm has been implemented to solve the Job Shop Scheduling Problem (JSSP). The chromosome design represents a feasible solution and meets all restrictions. The selection mechanism per tournament was used, a 95% reproduction based on partial pairing with two crossing points, a mixed strategy in the mutation stage combining the method of exchange and the method of investment using two random points in each machine and a percentage of progressive mutation between 2% to 5%. The results show that the algorithm must be executed with 100 individuals as population size and 500 generations for problems whose operation times are between 0 and 10 units of time and with 100 individuals and 1500 generations for problems between 0 and 100 units of time. The study shows that the implemented algorithm finds optimal solutions in the first case and highly competitive solutions in the second case. These are comparable with the results published in the literature that are generally responses to hybrid algorithms re-energized with other metaheuristics.
Título traducido de la contribución | Simple genetic algorithm to solve the Job Shop Scheduling Problem |
---|---|
Idioma original | Español |
Páginas (desde-hasta) | 299-313 |
Número de páginas | 15 |
Publicación | Informacion Tecnologica |
Volumen | 29 |
N.º | 5 |
DOI | |
Estado | Publicada - oct. 2018 |
Publicado de forma externa | Sí |
Palabras clave
- Assigning tasks
- Detailed planning
- Genetic algorithm
- Sequence of operations