In the present work, a heterogeneous breast model comprising of fat, glandular and muscle tissues along with a spherical tumor of 2 cm diameter is modelled. The efficacy of Radio Frequency Ablation (RFA) is mainly judged by the extent to which the tumor cells are damaged and hence the volume fraction of necrotic tissue (VFNT) is a response parameter. Several critical factors affect the volume of necrotic tissue, among which the blood perfusion rate, target tip temperature of electrode and location of tumor from the body core are considered as prime input parameters. This work focuses on the selection and the significance of tolerance of these critical factors using ‘cross-array design of experiment’ (CA-DOE) approach. The loose and tight ranges of tolerances of these factors are selected to form Taguchi’s inner and outer orthogonal array and experimental combination sets are prepared. In this study, the effect of each possible set of tolerance on the output parameter i.e. tumor ablation volume generated is analyzed. The different values of the ablation volume generated for each experiment are obtained from finite element analysis. The entire analysis is carried out using a bioheat transfer module coupled with electric currents module in COMSOL multiphysics software. Additionally, target tip temperature is controlled by the implementation of proportional-integral-derivative control strategy. The simulations results are validated with the available published experimental results. The simulation results are further analyzed using statistical methods, such as signal to noise ratio, analysis of mean, and analysis of variance. The presented analysis will be helpful to achieve robust and targeted values of ablation volume with least effort of maintaining process accuracy. The proposed methodology and results are useful for clinicians in the RFA.