Hybridization of Evolutionary and Swarm Intelligence Algorithms for improved performance

Abstract

This paper conducts the hybridization of Swarm intelligence and Evolutionary Algorithm for Continuous and Discrete optimization. Optimization is the process of selecting the best element by following some rules and criteria from some set of available alternatives. Function optimization means finding the best available value of some given objective function in a defined domain. In this work we have proposed an innovative approach, by hybridizing Genetic Algorithm (GA) and Swarm Intelligence Algorithm (SIA). In this paper work we have implemented one evolutionary programming based algorithm - Improved First Evolutionary Programming (IFEP) and one swarm intelligence algorithm - Ant Colony Optimization (ACO). We have also used Travelling Salesman Problem (TSP) as a discrete problem. We have implemented both GA and ACO also to solve the Travelling Salesman Problem. We have compared the result produced by IFEP and ACO for Continuous Optimization. From the comparative study we have found that ACO is the better among the two. We also have compared the result produced by GA and ACO for Discrete Optimization and from the comparative study we have found that ACO often works better. We have conducted some experiments to optimize the parameters of ACO and GA and the amount of exploration and exploitation needed for ACO to produce the best result. using the best found parameter we have implemented a hybrid of Genetic Algorithm and Swarm Intelligence Algorithm and tested it with different strategies. Then we have conducted a comparative study between the hybrid and two other conventional Genetic and Swarm Intelligence Algorithms to observe the performance of our proposed hybrid algorithm. In some cases we have observed better performance from our proposed hybrid algorithm.