Solution of the inverse bioheat transfer problem for the detection of tumors by genetic algorithms

Abstract

The problem of determining the size and location of a tumor situated underneath the skin by means of an inverse bioheat transfer analysis is considered in this article. The problem is posed by minimizing an error norm that considers the temperature information at the skin surface. Since Genetic Algorithms (GA’s) are powerful and versatile tools, a GA of steady–state type is implemented to solve the optimization problem originated from the inverse analysis. A finite element program based on the discretization of Pennes’ bioheat equation coupled with the Gmsh software is also developed to solve the set of direct problems required by the GA. A 2D numerical model is analyzed in order to demonstrate the effectiveness and robustness of the proposed approach.