Parameter estimation in a biological system using differential evolution algorithm

Abstract

Tumours generally involve high rates of metabolic heat generation and blood perfusion. In this paper, we present an inverse method involving the differential evolution algorithm for predicting the blood perfusion rate from the knowledge of transient temperature response of the skin. The interesting aspect of this work is to demonstrate that mere prediction of blood perfusion rate can characterize a tumorous tissue without any prior knowledge of the rate of metabolic heat generation. This is done by the incorporation of the initial temperature in the governing forward method itself and eliminating the metabolic heat generation from the pertinent expressions. Due to the nonhomogenous nature of biological tissues, the thermal relaxation time of such systems is considerably higher than other materials. Thus, unlike conventionally-studied Pennes bioheat transfer model, the present study addresses a non-Fourier heat conduction-based bioheat transfer model. The effect of random noise is also accounted in the present study and it is finally observed that the present work satisfactorily deciphers the malignant melanoma and other related subsurface abnormalities using a non-invasive inverse method aided by the skin’s transient thermal signatures.