A Study on Fitness Function with Parameter-free Genetic Algorithm for Non-destructive Estimation of Permittivity Distribution based on Capacitance and Electric Potential

Abstract

We have carried out experimental and numerical analyses for reconstruction of permittivity distribution using a pair of long distance small size electrodes and additional arc electrodes. A reconstruction method for permittivity distribution that uses a genetic algorithm and numerical analysis (finite element method) based on measured capacitance has been proposed and investigated. However, the conventional method had a problem in accuracy. In this paper, the reconstruction method has been improved by using a Parameter-free Genetic Algorithm(PfGA) based on measured capacitance and electric potential. In addition, we have evaluated the fitness function with PfGA.