Comparative analysis of chaotic variant of firefly algorithm, flower pollination algorithm and dragonfly algorithm for high dimension non-linear test functions

Abstract

Non-linear test functions are NP-Class problems. To solve them, Swarm Algorithms (SA) have been used in last two decades very effectively. In this work, three swarm based algorithms (i.e. Firefly Algorithm (FFA); Flower Pollination Algorithm (FPA) and Dragonfly Algorithm (DA)) have been used. Chaos is familiarized with swarm algorithm to improve their performance. As per our knowledge, most of the studies have applied chaos on one standard SA and compared it with other standard algorithm(s). No comparison has been shown among the chaotic variant of different algorithms. Comparison of Chaotic variants of FFA, FPA & DA with their standard algorithms has been performed using four high dimensions non-linear test functions on the basis of Mean fitness (i.e. P1) and convergence rate (i.e. P2). The results indicate that chaotic variant has performed better than standard and FFA evaluates best fitness for multi-modal function (i.e. f3 and f4).