Parallel hybrid evolutionary algorithm based on chaos-GA-PSO for SPICE model parameter extraction

SPICE model is one of the key technical connections between the integrated-circuit technology community and the design community. Design community requires accurate SPICE model parameters so as to make the difference between design spec and practical spec minimized as possible. To get accurate model parameters, optimization algorithms are used for parameter extraction. A parallel hybrid evolutionary algorithm based chaos-GA-PSO is presented for SPICE model parameter extraction, which gets leverage of the advantages from chaos algorithm, genetic algorithm, and particle swarm optimization, overcomes their respective disadvantages, passes good individuals among them, avoids local area optimization efficiently, and gets more accurate global optimized parameter extraction results. Also such algorithm based SPICE model parameter extraction architecture is presented with model convergence checking and derivate ability checking supported.