A Novel Parameter Optimization Metaheuristic: Human Habitation Behavior Based Optimization
Divya Jain, Mithlesh Arya, Varun Malik, S. Vikram Singh
Abstract
Parameters have great impact on the performance of an optimization algorithm. This paper concerns parameter tuning for metaheuristics which are stochastic optimization algorithms. A novel, fast, and very simple population-based metaheuristic namely Human Habitation Behavior-Based optimization (HHBO) is proposed in order to tune parameters of the metaheuristics used for the weight and bias optimization problem in feed-forward neural networks. The proposed algorithm is compared with other state-of-art algorithms, and results and analysis are presented. The results show the merits of HHBO for parameter tuning in comparison of other state-of-art algorithms.
Topics & Concepts
MetaheuristicComputer scienceMathematical optimizationMeta-optimizationParallel metaheuristicOptimization problemArtificial neural networkContinuous optimizationPopulationMulti-swarm optimizationAlgorithmArtificial intelligenceMathematicsDemographySociologyMetaheuristic Optimization Algorithms ResearchNeural Networks and ApplicationsEvolutionary Algorithms and Applications