Neural-Network Based Adaptation of Variation Operators’ Parameters for Metaheuristics

Abstract

The paper presents an idea of training an artificial neural network a relation between different parameters observed for a population in a metaheuristic algorithm. Then such trained network may be used for controlling other algorithms (if the network is trained in such way, that the knowledge gathered by it becomes agnostic regarding the problem). The paper focuses on showing the idea and also provides selected experimental results obtained after applying the proposed algorithm for solving popular benchmark problems in different dimensions.