Development of Hybrid Artificial Neural Network–Particle Swarm Optimization Model and Comparison of Genetic and Particle Swarm Algorithms for Optimization of Machining Fixture Layout

In this research paper, a methodology is proposed by combining taguchi’s parametric design, hybrid artificial neural network–particle swarm optimization (ANN–PSO) and evolutionary techniques to optimize the fixture layout by minimizing the maximum workpiece deformation on a 2D fixture workpiece syst...

Full description

Saved in:
Bibliographic Details
Published in:International journal of precision engineering and manufacturing Vol. 23; no. 12; pp. 1411 - 1430
Main Authors: Ramesh, M., Sundararaman, K. A., Sabareeswaran, M., Srinivasan, R.
Format: Journal Article
Language:English
Published: Seoul Korean Society for Precision Engineering 01.12.2022
Springer Nature B.V
Subjects:
Artificial neural networks
Deformation
End milling
Engineering
Finite element method
Genetic algorithms
Industrial and Production Engineering
Layouts
Materials Science
Neural networks
Optimization
Optimization models
Particle swarm optimization
Regular Paper
Response surface methodology
Scientific papers
Workpieces
Finite element method
Evolutionary techniques
Artificial neural network
Genetic algorithm
Particle swarm optimization
Fixture layout optimization
ISSN:2234-7593, 2005-4602
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this research paper, a methodology is proposed by combining taguchi’s parametric design, hybrid artificial neural network–particle swarm optimization (ANN–PSO) and evolutionary techniques to optimize the fixture layout by minimizing the maximum workpiece deformation on a 2D fixture workpiece system in end milling operation. Taguchi’s parametric design with five levels is utilized iteratively to estimate the potential range to place the fixture elements around the workpiece using the data obtained from finite element method. The hybrid ANN–PSO model is developed to predict the maximum workpiece deformation within the potential range in which PSO is utilized to optimize the weights and biases of the network. The diversity of data used for training the model is ensured by combining the experimental conditions of central composite design and Box Behnken design of response surface methodology. The developed model is tested using root mean square error, which exhibited better prediction accuracy. The hybrid ANN–PSO model is then optimized by genetic algorithm (GA) and PSO. The results clearly indicate that the PSO is capable of producing better fixture layouts with 0.1936% of superiority in solution quality than GA. Hence, the proposed approach is more viable to design the improved fixture layout with huge reduction in time and computational complexity.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2234-7593
2005-4602
DOI:10.1007/s12541-022-00698-z