A new optimization algorithm applied in electromagnetics — Maxwell’s equations derived optimization (MEDO)

Abstract

An algorithm of sequential systems of linear equations for nonlinear optimization problems with arbitrary initial point Science in China Series A-Mathematics 40, 561 (1997); Numerical optimization method for HJI equations derived from robust receding horizon control schemes and controller design SCIENCE CHINA Information Sciences 55, 214 (2012); A new algorithm for the problem of robust single objective optimization

Adaptive Maxwell's equations derived optimization and its application in antenna array synthesis

“…MEDO is an optimization algorithm, which evolves from the Law of Conservation of Charge and Faraday's Law, and shows strong potential in the applications of electromagnetics [1]. The algorithm is inspired by summarizing the behavior of the charged conductors in time-varying electromagnetic fields.…”

“…They refer to the gravitational acceleration, the magnetic flux density, the per-unit-lengthweight of the conductor, the resistant of the main branch, and the loop inductance of current loop 1, respectively. Reference has proved that the individuals can converge to the optimal location of the objective function under the appropriate parameter settings [1].…”

“…One typical benchmark is calculated by MEDO, with each parameter varies independently. Other parameters take the best values suggested in [1]. The curves of the convergence errors that change with the variation of each parameter are plotted in Fifure 2.…”

“…In electromagnetics and communications, the optimization problems are usually high-dimensional, multimodal and complex, which are challenging for the optimization algorithms. In order to solve the optimization problems in the electromagnetics and communications more efficiently, Maxwell's Equations Derived Optimization (MEDO) is proposed, which is one of the typical representatives of the third type of the algorithms [1].…”

“…Compared with other popular optimization algorithms such as Gradient Descent (GD) [2], Particle Swarm Optimization (PSO) [3,4], Differential Evolution (DE) [5], and Adaptive Wind Driven Optimization (AWDO) [6], MEDO can attain more accurate and more stable optimization results for both unimodal functions and multimodal functions [1]. However, the performance of MEDO is sensitive to the parameters, which need to be tuned manually for different problems.…”