Outlander Algorithm Based on Integrated Aggressive Selection

The new algorithm called “Outlander” that promotes better individuals, allows convergence and maintains exploration abilities.

Numerous algorithms are utilized in optimization problems. One of the most commonly used methods to find optimum points of a given function is Genetic Algorithms, which stochastically select individuals from the population. The aim of genetic algorithms is to gradually approximate to the optimum points by choosing the better individuals in each iteration. Thus, having a good selection method is significantly important issue in genetic algorithms. In this paper, a new selection method, which is an improved version of Integrated Aggressive Selection Method, is introduced. The performance of the new method is compared with four methods that were previously proposed by us; Aggressive, Integrated Aggressive, Non-Aggressive and Integrated Non-Aggressive selection methods, and with the most commonly used standard selection methods; Roulette Wheel, Linear Ranking and Tournament by utilizing variety of benchmark functions. It is observed that newly proposed "Outlander" algorithm delivers the overall best performance results comparing to the other selection methods for both unimodal and multimodal optimization problems.