Initialization method of genetic algorithm based on improved clustering algorithm
Authors:
Open Access
- 9 July 2022
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 447-450
Abstract
In order to improve the search ability, the convergence performance and algorithm results of genetic algorithm, we propose a population initialization method of genetic algorithm based on improved k-means algorithm. Firstly, the initial individuals walking evenly in the decision space are created by uniform design method, and then the initial individuals are processed by local search such as step-by-step search. Secondly, through the similarity degree between the individuals, several main search spaces are aggregated. Based on the search space, the individuals needed by the final initial population are clustered by K-means method.Keywords
This publication has 3 references indexed in Scilit:
- Improved K-means algorithm based on density CanopyKnowledge-Based Systems, 2018
- Sequence-Based Deterministic Initialization for Evolutionary AlgorithmsIEEE Transactions on Cybernetics, 2016
- Cluster-Based Population Initialization for differential evolution frameworksInformation Sciences, 2015