Efficient Differential Grouping Method for Large-scale Constrained Optimization

• Autores: Paz, Fabiola; Leguizamón, Guillermo; Mezura Montes, Efrén
• Año de publicación: 2024
• Idioma: inglés
• Tipo de recurso: documento de conferencia
• Materia: Ciencias Informáticas; Differential Grouping; Interaction Between Variables; Decomposition Methods; Cooperative Co-evolution Algorithms; Large-scale Constrained Optimization
• Institución: Universidad Nacional de La Plata

Solving large-scale constrained optimization problems (LSCOP) using cooperative co-evolutionary (CC) algorithms has received considerable attention in recent years. One of the most critical challenges of CC algorithms is the decomposition of the original problem. Currently, decomposition methods for solving LSCOPs often generate large subcomponents, poor accuracy, and consume significant computational resources. This affects the optimization performance. In this work, we present a decomposition method that aims to form efficient subcomponents that feature high accuracy, good size, and low computational resource cost. The proposal is evaluated on a set of benchmark functions subject to constraints up to 1,000 dimensions widely used in the literature, and compared with other state-of-the-art methods. The results demonstrate that our proposal performs the decomposition efficiently and succeeds in reducing the computational cost, making it a valuable contribution to the field of optimization.
Red de Universidades con Carreras en Informática