Optimización basada en colonias de hormigas en dos etapas
Fecha
2007-06-10
Autores
Puris Cáceres, Amílkar Yudier
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas. Facultad de Matemática, Física y Computación. Departamento de Computación
Resumen
La metaheurística “Optimización basada en colonias de hormigas” (Ant Colony Optimization,
ACO) es uno de los nuevos paradigmas que permite la solución de problemas combinatorios del tipo
NP-Hard. En este trabajo se presenta un nuevo modelo basado en el comportamiento de las
hormigas nombrado “Optimización basada en colonias de hormigas en dos Etapas”( Two Steage Ant
Colony Optimization, TS-ACO). La nueva estrategia de exploración que presenta este modelo
propone hacer una división en dos del espacio de búsqueda, donde en la primera etapa solo una parte
de las hormigas van a solucionar un subproblema de tamaño inferior al problema original, estas
subsoluciones encontradas servirán de punto de partida para que en la segunda etapa las hormigas
restantes de la colonia busquen soluciones al problema en general, haciéndose de esta forma más
cooperativo el trabajo de estos agentes. El nuevo modelo muestra un mayor rendimiento en cuanto
al tiempo de ejecución y la calidad de las soluciones encontradas. Para probar la nueva estrategia se
escogieron los algoritmos Ant Colony System (ACS) y Ant System (AS) para darle solución a tres
problemas de optimización combinatoria, el Viajante de Comercio (TSP), el problema de
Secuenciación de Tareas (JSSP) y el problema de Cubrimiento de Conjuntos (SCP).
The Ant Colony Optimization (ACO) meta-heuristic is one of the new paradigms that allow solving NP-Hard combinatorial problems. In this study, a new ant-behavior-based model named “Two Stage Ant Colony Optimization” (TS-ACO) is formally introduced. The fresh exploration strategy splits the search space into two parts. During the first stage, only a subset of the available ants is intended to solve a sub-problem of the original problem. The solutions found herein will serve as the starting point for the remaining ants of the colony to perform at the second stage, leading to a tighter cooperation between those agents. The proposed model exhibits a higher performance in terms of the execution time and the quality of the solutions yielded. In order to test our approach, the Ant Colony System (ACS) and Ant System (AS) algorithms were chosen as benchmarks for being applied to three well-known combinatorial optimization problems, namely the Traveling Salesman Problem (TSP), the Job Shop Scheduling Problem (JSSP) and the Set Covering Problem (SCP).
The Ant Colony Optimization (ACO) meta-heuristic is one of the new paradigms that allow solving NP-Hard combinatorial problems. In this study, a new ant-behavior-based model named “Two Stage Ant Colony Optimization” (TS-ACO) is formally introduced. The fresh exploration strategy splits the search space into two parts. During the first stage, only a subset of the available ants is intended to solve a sub-problem of the original problem. The solutions found herein will serve as the starting point for the remaining ants of the colony to perform at the second stage, leading to a tighter cooperation between those agents. The proposed model exhibits a higher performance in terms of the execution time and the quality of the solutions yielded. In order to test our approach, the Ant Colony System (ACS) and Ant System (AS) algorithms were chosen as benchmarks for being applied to three well-known combinatorial optimization problems, namely the Traveling Salesman Problem (TSP), the Job Shop Scheduling Problem (JSSP) and the Set Covering Problem (SCP).
Descripción
Palabras clave
Colonias de Hormigas, Ant Colony System, Ant System, Viajante de Comercio, Secuenciación de Tareas, Cubrimiento de Conjuntos