Performance comparison of GA, PSO, and DE approaches in estimating low atmospheric refractivity profiles

Abstract

Particles swarm optimization (PSO) and differential evolution (DE) algorithms based on optimization are employed to estimate low atmospheric refractivity profiles from radar sea clutter. Low atmospheric refractivity profiles are modeled as evaporation ducts. The objective functions, which are used to evaluate the fit of simulated and measured power in estimation procedures, are also investigated at different frequencies such as L-, S-, C-and Xfrequency at 10 m/s wind speeds. The results show that all the objective functions are multi-peak functions. The Adjusted Barton Model of radar cross section (RCS) is adopted. PSO and DE algorithms are compared with genetic algorithm (GA) by 200 Monte Carlo simulation estimations. Simulation results indicate that DE has the best global search ability, and PSO has the highest success probability. According to the statistical results, PSO algorithm with the population size 30 is the appropriate way for evaporation duct estimation.

Easy Particle Swarm Optimization for Nonlinear Constrained Optimization Problems

“…ng p is denoted as the best position for group n in the swarm. The formula of the local search PSO can be changed to Equations ( 10)- (11).…”

“…[10] have used PSO, DE, and genetic algorithm (GA) altogether to estimate low atmospheric refractivity profiles from radar sea clutter. Such studies essentially recognized PSO's superiority compared with other algorithms for NO problems [9] [11].…”

Solving Many-Objective Car Sequencing Problems on Two-Sided Assembly Lines Using an Adaptive Differential Evolutionary Algorithm

“…one point crossover, which is normally used in the GA. However, the GA was outperformed by the differential evolution (DE) algorithm in terms of global search ability [44,45]. Li and Zhang [46], Li and Zhang [47] proposed a multi-objective differential evolution based decomposition (MODE/D) in which the genetic operator was replaced by the differential operators and revealed that the MODE/D performed better than several other MOEAs on many test problems.…”