Nature-Inspired Algorithms and Applications

6Firefly Algorithm

Anupriya Jain, Seema Sharma* and Sachin Sharma

Manav Rachna International Institute of Research and Studies, Faridabad, India

Abstract

Data science community has been in continuous search for inspiration from nature and humans to solve various real-life problems. Machine learning itself is based on the way humans learn. Meanwhile, nature-inspired algorithms have gained popularity to solve complex optimization problems. These are termed as biology/nature-inspired algorithms. Researchers have found swarm intelligence algorithms that work on the principle of a swarm of bees, ants, etc. One of the most popular algorithms that use the behavior of fireflies, their light flashing patterns in order to find a global optimal solution is Firefly Algorithm (FA). In this chapter, we present the working principle of FA in detail with the algorithm explained and its implementation ready for the reference. In recent years, there has been the introduction of variants of FA to accommodate new problems. The hybrid or modified models have tremendously improved the performance of a standard FA. These special cases and applications of this meta-heuristic problem are discussed in detail. We also determine the need for FA over swarm optimization and why does FA works so well.

Keywords: Firefly Algorithm (FA), machine learning, swarm optimization, meta-heuristic problem

6.1 Introduction

Optimization refers to finding the optimal, meaning the “best” or “most suitable” solution ...