2016/02/09 by Shayan Poursoltan, Frank Neumann, Poursoltan, Shayan +1
Biochemistry, Genetics and Molecular Biology · Computer Science · Engineering · #DNA and Biological Computing #Evolutionary Algorithms and Applications #FOS: Computer and information sciences #Modular Robots and Swarm Intelligence #Neural Networks and Applications #Neural and Evolutionary Computing (cs.NE)
paper · pdf · doi:10.48550/arxiv.1602.02862
openalex publication_date 2016/02/09 · openalex created_date 2022/10/01 · openalex updated_date 2026/07/28
With this paper, we contribute to the growing research area of feature-based\nanalysis of bio-inspired computing. In this research area, problem instances\nare classified according to different features of the underlying problem in\nterms of their difficulty of being solved by a particular algorithm. We\ninvestigate the impact of different sets of evolved instances for building\nprediction models in the area of algorithm selection. Building on the work of\nPoursoltan and Neumann [11,10], we consider how evolved instances can be used\nto predict the best performing algorithm for constrained continuous\noptimisation from a set of bio-inspired computing methods, namely high\nperforming variants of differential evolution, particle swarm optimization, and\nevolution strategies. Our experimental results show that instances evolved with\na multi-objective approach in combination with random instances of the\nunderlying problem allow to build a model that accurately predicts the best\nperforming algorithm for a wide range of problem instances.\n