Special Relativity Search: A novel metaheuristic method based on special relativity physics

• Published in: Knowledge-based systems Vol. 257; p. 109484
• Main Authors: Goodarzimehr, Vahid, Shojaee, Saeed, Hamzehei-Javaran, Saleh, Talatahari, Siamak
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 05-12-2022
• Subjects: Global optimal; Magnetic field; Metaheuristic; Optimization; Special Relativity Search; Metaheuristic; Special Relativity Search; Magnetic field; Optimization; Global optimal
• ISSN: 0950-7051; 1872-7409
• DOI: 10.1016/j.knosys.2022.109484

In this work, a novel metaheuristic optimization algorithm called Special Relativity Search (SRS) is proposed. The SRS is inspired by the interaction of particles in an electromagnetic field. Particle interactions are calculated using the Lorentz force, and the equation of motion is developed using angular frequency. The magnetic force between particles is perpendicular to the velocity of charged particles and the magnetic field, which causes particles to move in a circular trajectory. In this method, for the first time, the theory of special relativity physics is utilized to determine the coordinates of charged particles in each rotation. The SRS main step equation is developed using the two phenomena of length contraction and time dilation. Charged particles are members of the initial population that is randomly generated, and their charge is determined based on their fitness. To show the efficiency and robustness of the SRS in solving optimization problems, 83 benchmark functions, which are a wide range of mathematical problems, are selected and optimized based on the values of Best, Mean, Median, and Standard Deviation (SD). The statistical test of the Wilcoxon Signed Ranks (WSR) is performed to fairly compare the results of this new method with other popular metaheuristic algorithms. The test results show that in most cases SRS is superior to other methods from the state-of-the-art. The results of evaluated optimization problems show that SRS is more effective and efficient compared to some other well-known metaheuristic methods in solving optimization problems.