.png){kind=icon}
.png){kind=icon}
Return
An adaptive differential evolution algorithm with population size reduction strategy for unconstrained optimization problem
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2023.110209.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The differential evolution (DE) algorithm is a heuristic random search algorithm that optimizes the problem based on population evolution. It has been widely studied for its advantages of fewer control parameters, fast convergence speed, and strong robustness. Here, an Adaptive Guided Differential Evolution algorithm on Mutation, Parameter and Population (AGDE-MPP) is proposed, which has improved the DE algorithm by adopting a new mutation scheme, a parameter adaptation scheme, and a non-linear population size reduction strategy. The new mutation scheme uses two ordered difference vectors to perturb the mutation direction of the base vector. The parameter adaptation scheme generates the mutation factor by the Cauchy distribution with the mean of the Lehmer mean value from the historical successful mutation factor pool. The new non-linear population size reduction strategy adopts a hyperbolic tangent function curve to control the population size. This adaptive algorithm can balance the exploration and exploitation capabilities at different stages of evolution and obtain solutions with high accuracy and fast convergence speed. The performance of AGDE-MPP and several state-of-the-art algorithms were evaluated on CEC2013 and CEC2017 test suites for 10, 30, and 50 dimensions. The experimental results show that the performance of AGDE-MPP is superior to almost compared algorithm and has strong competitiveness.& COPY; 2023 Elsevier B.V. All rights reserved.
Keywords:
Differential evolution algorithm
Novel mutation scheme
Parameter adaptation scheme
Population size reduction
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
View more organizations