Global Annealing Strategy: Evolutionary Circuit Design Made Efficiently

Premature convergence or fail on convergence within a limited period of time are difficulties in the field of evolutionary circuit design and the problems are related to the selection pressure of algorithms and population diversity. In this paper, we present a global annealing strategy for evolutionary circuit design, which is a new attempt to combine genetic algorithm and simulated annealing algorithm. Finally, an inverting amplifier is taken as a test case and a series of contrast experiments are conducted under the same experimental execution environment. The results reveal that the strategy presented performs better on capturing the desirable circuit with efficient search ability and it processes better convergence performance, namely, higher success rate within a limited evolutionary generation.