Search

Evolutionary algorithms, such as genetic algorithms and particle swarm optimization, are optimization and search algorithms inspired by the process of natural selection. These algorithms evolve a population of potential solutions over multiple generations to find optimal solutions to a given problem. In many optimization problems, there can be multiple optima, which may be isolated from each other in the search space. The evolutionary algorithm might converge prematurely to a single optimum, missing other potentially better solutions. Niching strategies aim to address this issue by encouraging diversity within the population. This work presents novel variants of three popular evolutionary algorithms Genetic Algorithm, Binary Differential Evolutionary Algorithm and Binary Particle Swarm Optimization Algorithm with niching using multiple populations to obtain multiple optima. These algorithms have been applied on a set of 71 fiducial ECG features to obtain multiple optimized subsets. The optimization criteria are to maximize accuracy of SVM classifier and at same time minimize the number of features in the subset. 100 unique optimized subsets are obtained with each algorithm. [ABSTRACT FROM AUTHOR]