Performance Evolution of Optimization Techniques for Mathematical Benchmark Functions

Abstract

This paper demonstrates several optimization techniques which comprise Genetic algorithm (GA), Ant Colony optimization (ACO) and Particle swarm optimization (PSO). The proposed paper enforces the concept of artificial intelligence to detect minima / maxima by applying set of mathematical benchmark functions. For Optimization technique, artificial intelligence used which comprise of several algorithms like Particle swarm optimization, genetic algorithm, ant colony optimization, neural network and fuzzy system. The proposed work will use particle swarm intelligence and genetic intelligence. This paper bestows comparison between PSO and GA according to performance. In random search algorithm, GA cannot detect the global optimization solution. The benchmark functions used under these algorithms are Rosenbrock, Griewank, Ackley and Sphere. They have multiple local minima / maxima and single global minima / maxima. Neural network has propensity to strikes at local minima / maxima. Result demonstrates that discomfort of neural network is thoroughly segregated by particle swarm intelligence and genetic algorithm