Tribes Optimization Algorithm Applied to the Loney's Solenoid

Abstract

Optimization problems in electromagnetics often exhibit multiple local optima, especially in the case of noisy objective functions. Particle swarm optimization (PSO) is a global search strategy which can efficiently handle arbitrary optimization problems in electromagnetics. The standard version of PSO is simple but needs the setting of some control parameters (swarm size, neighborhoods, inertia weight, cognitive, and social parameters). This paper evaluates an adaptive swarm intelligence approach called TRIBES that does not have this drawback, i.e., TRIBES is a parameter-free PSO algorithm. Numerical results for Loney's solenoid benchmark problem demonstrate the applicability and efficiency of the TRIBES algorithm.