The Use of Genetic Algorithms for Optimizing the Regularized Solutions of the Ill-Posed Problems

Numerical solution of ill-posed problems is often accomplished by regularization method, such as Tikhonove method, truncated singular-value decomposition (TSVD). Due to large noises condition, those conventional regularization methods cannot provide available solutions for ill-posed problem. Considering that genetic algorithms (GA) is a stochastic optimization technique which may be useful for optimizing the regularized solutions. Computing the epicardial potentials from the body surface potentials constitutes one form of the ill-posed inverse problems of electrocardiography (ECG), which is considered as an example to illustrate the performance of GA when applied to optimize the regularized solutions. The result suggests that the GA may be a good scheme for optimizing the regularized solutions in solving the inverse ECG problem.