Application of swarm and evolutionary techniques to neural network training and partitioning combinational CMOS

Swarm Intelligence...involves a population of simple agents interacting locally with one another and with their environment, leading to the emergence of global behavior. Particle Swarm Optimization (PSO) and quantum-inspired evolutionary algorithm (QEA) are two such evolutionary computation techniques. Much simulation work exists in the literature on evolutionary algorithms performing well on optimization problems, but few hardware applications are reported. This thesis aims at finding suitable algorithms for hardware implementation on digital processor. Thus, Binary PSO, QEA and Improved QEA are investigated on basic nonlinear function approximation problems --Abstract, page iii