Enhanced Special Relativity Search Algorithm-Based Lorentz Force for Structural Optimization

Abstract

Optimization of real-world problems is an important challenge that researchers face due to the discrete nature of the search space, multiple local optimum, the large dimension of the problem, and computation costs. The Special Relativity Search (SRS) algorithm is one of the newest metaheuristic methods that has recently been developed. In this work, the Enhanced SRS (ESRS) algorithm has been developed based on the Lorentz Force coefficient to solve this class of problems. The SRS is a single-objective algorithm based on swarm intelligence inspired by the physics of special relativity. The cause of movement between particles in magnetic space is the Lorentz Force. Due to the type of particle charge, this force repels or attracts particles. SRS performance is sensitive to the Lorentz Force. Developing an empirical equation can enhance the performance of the algorithm. An empirical equation for simulating the Lorentz Force Coefficient is proposed as LC. LC decreases proportionally to the number of different iterations. This equation has a constant coefficient to control the movement step. This constant is more accurate for small values than for large values. Several structural problems with discrete and continuous variables have been investigated to evaluate the performance of the proposed algorithm. The objective function is the weight of structural elements that are under loading and must be reduced to the minimum value while observing the constraints of the problem. ESRS results have been compared with the original SRS and several state-of-the-art algorithms. The results show that ESRS has obtained the best optimal weight with the least computational effort and with high accuracy.