Comparison of Quantum-inspired Evolutionary Algorithm and Particle Swarm Optimization for Neural Network Training
Neural Networks (NN) are known to be unversal approximators for any non-linear function. Training algorithms are critical when neural networks are applied to high speed applications with complex nonlinearities. The performance of training algorithm varies depending on the type of neural network. Quantum-Inspired Evolutionary Algorithm (QEA) is proposed as a novel technique for NN training, which is based on the concept and principles of quantum computing. QEA utilizes the concepts of superposition of states to do parallel processing on all the possible solutions. To demonstrate its effectiveness, studies are carried out for training feedforward neural networks (FFNNs) to learn time-independent and time-dependent nonlinear functions. The training with QEA has been compared with Binary Particle Swarm Optimization (BPSO) algorithm in this paper.
G. Singhal and G. K. Venayagamoorthy, "Comparison of Quantum-inspired Evolutionary Algorithm and Particle Swarm Optimization for Neural Network Training," Conference on Neuro-Computing and Evolving Intelligence, Auckland, New Zealand, Knowledge Engineering and Discovery Research Institute, Dec 2004.
Electrical and Computer Engineering
National Science Foundation (U.S.)
Keywords and Phrases
Neural Networks; Particle Swarm Optimization; Quantum Evolutionary Algorithm
Article - Conference proceedings
© 2004 Knowledge Engineering and Discovery Research Institute, All rights reserved.
This document is currently not available here.