An empirical study of impact of crossover operators on the performance of non-binary genetic algorithm based neural approaches for classification
Computers & Operations ResearchPublished 11 February 2003
Parag C. Pendharkar, James A. Rodger
Citations72
SJR quartileQ1
SJR score1.60
SNIP2.02
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
The results indicate that arithmetic crossover operator may be a suitable crossover operator for GA based ANN.
Abstract
We study the performance of genetic algorithm (GA) based artificial neural network (ANN) for different crossover operators. We use simulated and real life data to test the performance of GA-based ANN. Our results indicate that arithmetic crossover operator may be a suitable crossover operator for GA based ANN. Scope and purpose Genetic algorithm based artificial neural networks are used in several classification and forecasting applications. Among several genetic algorithm design operators, crossover plays an important role for convergence to the global heuristic solution. Several crossover operators exist, and selection of a crossover operator is an important design issue confronted by most researchers. The current study investigates the impact of different crossover operators on the performance of genetic algorithm based artificial neural networks.
Keywords
Computer Science
Choice Reviews OnlineGenetic algorithms in search, optimization, and machine learning
49,278 Citations1989
This book brings together the computer techniques, mathematical tools, and research results that will enable both students and practitioners to apply genetic algorithms to problems in many fields.
Foundations of genetic algorithmsA Comparative Analysis of Selection Schemes Used in Genetic Algorithms
2,400 Citations1991David E. Goldberg, Kalyanmoy Deb
A number of selection schemes commonly used in modern genetic algorithms are compared on the basis of solutions to deterministic difference or differential equations, verified through computer simulations to provide convenient approximate or exact solutions and useful convergence time and growth ratio estimates.
Journal of Banking & FinanceZETATM analysis A new model to identify bankruptcy risk of corporations
1,818 Citations1977Edward I. Altman, R. G. Haldeman +1 more
Foundations of genetic algorithmsGenetic Algorithms for Real Parameter Optimization
989 Citations1991Alden H. Wright
It is shown that k-point crossover can be viewed as a crossover operation on the vector of parameters plus perturbations of some of the parameters, which suggests a genetic algorithm that uses real parameter vectors as chromosomes, real parameters as genes, and real numbers as alleles.
Evolutionary ComputationForming Neural Networks Through Efficient and Adaptive Coevolution
307 Citations1997David E. Moriarty, Risto Miikkulainen
The symbiotic adaptive neuroevolution system coevolves a population of neurons that cooperate to form a functioning neural network to be more efficient and more adaptive and to maintain higher levels of diversity than the more common network-based population approaches.
IEEE Transactions on Evolutionary ComputationEvolutionary algorithms and gradient search: similarities and differences
277 Citations1998R. Salomon
To what extent a hybrid method, the evolutionary-gradient-search procedure, can be used beneficially in the field of continuous parameter optimization is explored.
European Journal of Operational ResearchOptimization of neural networks: A comparative analysis of the genetic algorithm and simulated annealing
233 Citations1999Randall S. Sexton, Robert E. Dorsey +1 more
This paper examines two well known global search techniques, Simulated Annealing and the Genetic Algorithm, and compares their performance, and a Monte Carlo study was conducted in order to test the appropriateness of theseglobal search techniques for optimizing neural networks.
Decision SciencesMinimizing Misclassifications in Linear Discriminant Analysis*
140 Citations1990Gary J. Kœhler, Ş. Selçuk Erengüç
Decision SciencesFOUR APPROACHES TO THE CLASSIFICATION PROBLEM IN DISCRIMINANT ANALYSIS: AN EXPERIMENTAL STUDY*
119 Citations1988Erich A. Joachimsthaler, Antonie Stam
Management ScienceUsing Feature Construction to Improve the Performance of Neural Networks
99 Citations1998Selwyn Piramuthu, Harish Ragavan +1 more
It is shown that the degree of difficulty in neural network learning is inherent in the given set of training examples, a technique for measuring such learning difficulty is proposed, and a feature construction methodology is developed that helps transform the training data so that both the learning speed and classification accuracy of neural network algorithms are improved.
Decision SciencesInductive, Evolutionary, and Neural Computing Techniques for Discrimination: A Comparative Study*
79 Citations1998Siddhartha Bhattacharyya, Parag C. Pendharkar
Simulated data is used to examine how the different learning techniques perform with respect to certain data distribution characteristics and helps relate the findings across a range of discrimination techniques.
OmegaNeural networks: a need for caution
72 Citations1997Bruce Curry, Phillip L. Morgan
It is suggested that the now traditional method of backpropagation (BP) may not be the most appropriate basis for learning, based on the known deficiencies of gradient descent methods.
Foundations of genetic algorithmsAn Extension To the Theory of Convergence and a Proof of the Time Complexity of Genetic Algorithms
56 Citations1991C. A. Ankenbrandt
The standard recurrence relation for binary and nonbinary genetic algorithms is solved using an inductive proof and a proof using linear finite difference equations, and the resulting equations define the relationship between estimated run time, problem size, fitness ratio and the complexity of the domain specific evaluation function.
Decision SciencesAn Efficient Optimal Solution Algorithm for the Classification Problem
50 Citations1991William J. Banks, P.L. Abad
A new LP-based algorithm is developed to determine the optimal solution to the two-group classification problem and is efficient when compared to current research in a simulation study.
European Journal of Operational ResearchNew LP based heuristics for the classification problem
49 Citations1993P.L. Abad, William J. Banks
Three heuristic procedures for the two group discriminant problem are developed that minimize the total expected cost of misclassification in the training sample and to be efficient in terms of computational effort when compared to existing research.
Annals of Operations ResearchNonlinear programming and genetic search application for production scheduling in coal mines
49 Citations2000Parag C. Pendharkar, James A. Rodger
A hypothetical model developed and tested using genetic search for nonlinear optimization indicates that the model has potential for decision support in coal mines.
European Journal of Operational ResearchA computational study on the performance of artificial neural networks under changing structural design and data distribution
35 Citations2002Parag C. Pendharkar
The results of the study indicate that additive noise, size, and data distribution characteristics play an important role in learning, reliability and predictive ability of ANNs.
INFORMS Journal on ComputingLinear Discriminant Functions Determined by Genetic Search
34 Citations1991Gary J. Kœhler
A genetic algorithm that determines linear discriminant functions is studied that routinely found the optimal solution while taking significantly less time than exact algorithms which must both find and prove optimality.
European Journal of Operational ResearchTraining neural networks with the GRG2 nonlinear optimizer
33 Citations1993Ming S. Hung, James W. Denton
This paper compares the use of a general-purpose nonlinear optimizer, GRG2, with back propagation in training neural networks and found thatGRG2 not only found solutions much faster, but also found much better solutions.
European Journal of Operational ResearchA combinatorial approach to the classification problem
31 Citations1999Nicola Yanev, Stefan Balev
This work studies the two-group classification problem which involves classifying an observation into one of two groups based on its attributes and presents exact and heuristic algorithms for solving the problem.
OmegaAn empirical study of design and testing of hybrid evolutionary–neural approach for classification
27 Citations2001Parag C. Pendharkar
Using a set of simulated and real life data sets, it is illustrated that the proposed hybrid approach fares well, both in training and holdout samples, when compared to the traditional back-propagation artificial neural network (ANN) and a genetic algorithm-based artificial Neural Network (GA-ANN).