Performance Analysis of Immune Plasma Algorithm with Different Donor-Receiver Configurations

Abstract

Immune Plasma Algorithm (IPA) is a novel meta-heuristic algorithm inspired by immune plasma transfer treatment. Many metaheuristic algorithms are used for solving complex optimization problems, but their performance is mostly inspected on problems with 30 dimensions. Nowadays we are dealing with far more complex systems that require solving high-dimensional optimization problems with over 50 dimensions whereas performance of meta-heuristic algorithms for high-dimensional problems is mostly unexamined. So to overcome this problem, in this study, performance of IPA on solving high-dimensional problems is investigated. In this case, it is used to solve five well-known benchmark optimization problems with 100 dimensions. In this work, Immune Plasma Algorithm (IPA) is used for solving Sphere, Quartic, Rastrigin, Ackley and Griewank functions. It is compared with some other stateof-the-art meta-heuristic algorithms. Experimental results demonstrate that IPA outperforms these algorithms in finding best objective values, and has best standard deviation, and best mean value for most of the tested optimization problems.