An efficient differential evolution algorithm with approximate fitness functions using neural networks

Authors:
Yi-Shou Wang;Yan-Jun Shi;Ben-Xian Yue;Hong-Fei Teng
Affiliations:
School of Aeronautics and Astronautics, Faculty of Vehicle Engineering and Mechanics, State Key Laboratory of Structural Analysis for Industrial Equipment, Dalian University of Technology, Dalian, ...;School of Mechanical Engineering, Dalian University of Technology, Dalian, P.R. China;Department of Computer Science and Engineering, Dalian University of Technology, Dalian, P.R. China;School of Mechanical Engineering, Dalian University of Technology, Dalian, P.R. China and Department of Computer Science and Engineering, Dalian University of Technology, Dalian, P.R. China
Venue:
AICI'10 Proceedings of the 2010 international conference on Artificial intelligence and computational intelligence: Part II
Year:
2010

Citing 9
Cited 0

Fundamentals of Artificial Neural Networks

Fundamentals of Artificial Neural Networks
Differential Evolution – A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces

Journal of Global Optimization
A comprehensive survey of fitness approximation in evolutionary computation

Soft Computing - A Fusion of Foundations, Methodologies and Applications
Combating user fatigue in iGAs: partial ordering, support vector machines, and synthetic fitness

GECCO '05 Proceedings of the 7th annual conference on Genetic and evolutionary computation
Memetic algorithm using multi-surrogates for computationally expensive optimization problems

Soft Computing - A Fusion of Foundations, Methodologies and Applications
Advances in Differential Evolution

Advances in Differential Evolution
System design by constraint adaptation and differential evolution

IEEE Transactions on Evolutionary Computation
A framework for evolutionary optimization with approximate fitnessfunctions

IEEE Transactions on Evolutionary Computation
Local function approximation in evolutionary algorithms for the optimization of costly functions

IEEE Transactions on Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

We develop an efficient differential evolution (DE) with neural networks-based approximating technique for computationally expensive problems, called DE-ANN hereinafter. We employ multilayer feedforward ANN to approximate the original problems for reducing the numbers of costly problems in DE. We also implement a fast training algorithm whose data samples use the population of DE. In the evolution process of DE, we combine the individual-based and generation-based methods for approximate model control. We compared the proposed algorithm with the conventional DE on three benchmark test functions. The experimental results showed that DE-ANN had capacity to be employed to deal with the computationally demanding real-world problems.