An adaptive differential evolution algorithm and its performance on real world optimization problems

Abstract

Real world optimization problems are challenging as they often involve a large number of variables and highly nonlinear constraints and objective functions. While a number of efficient optimization algorithms and numerous mathematical benchmark test functions have been introduced in recent years, the performance of such algorithms have rarely been studied across a range of real world optimization problems. In this paper, we introduce an improved adaptive differential evolution (DE) algorithm and report its performance on the newly proposed real world optimization problems. The proposed differential evolution algorithm incorporates adaptive parameter control strategies; a center based differential exponential crossover and hybridization with local search to improve its efficiency. While comprehensive results of other algorithms on the test problems are unavailable at this stage, our preliminary comparison with published results indicates promising performance of the proposed DE across the range of problems.

A Differential Evolution Algorithm with Constraint Sequencing

“…A population size of 100 has been used with F = 0.5 and a CR set of CR = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0} [16]. One can observe from table V that the proposed algorithm required less number of function evaluations for problems g01, g02, g04, g09, g10, g18 and g24 as compared to the methods proposed by Takahama and Sakai [8], Brest [12] and, Zavala [13].…”

Variations of biogeography-based optimization and Markov analysis

“…lems. We use the adaptive DE proposed by [2], where control parameter settings are gradually adapted according to the learning progress, and which uses a center based differential exponential crossover and incorporates local search to improve its efficiency. We recognize that there are many other evolutionary algorithms, including the estimation of distribution algorithm (EDA), evolutionary strategy (ES), ant colony optimization (ACO), and their variants, which may provide better optimization performance than the algorithms we use in this paper.…”

Differential evolution with a mix of Constraint Consenus methods for solving a real-world Optimization Problem

“…In this subsection, DEwMCC is compared of-the-art algorithms, in particular (i) Diffe with Multiple Strategies (SAMODE) [26] differential evolution algorithm (IDE) [27]. W here that DEwMCC used the same num evaluation (FEs) (150K) as used in SAMOD detailed results are shown in Table II.…”