Quantum-inspired Evolutionary Algorithms and Binary Particle Swarm Optimization for Training MLP and SRN Neural Networks

This paper presents a comparison of two machine learning methods inspired by nano-scale and macro-scale natural processes and related to distributed intelligence, namely Quantum—Inspired Evolutionary Algorithm (QEA) and Binary Particle Swarm Optimization (BPSO). QEA is based on the concepts and principles of Quantum Computing, such as a quantum bit (Q-bit) and superposition of states. QEA uses a Q-bit for the probabilistic representation and a Q-bit individual as a string of Q-bits. A modified QEA with less memory requirements is also presented. The effectiveness of these algorithms in binary search space are compared for training neural networks. Results are presented for Multilayer Perceptrons (MLPs) and Simultaneous Recurrent Neural Networks (SRNs). For neural networks trained on complex nonlinear functions, the QEA based algorithms achieve convergence faster than BPSO