This work presents the results of the multi objective geometric optimization of a phase change material (PCM) based cylindrical heat sink with thermal conductivity enhancers (TCEs) in the form of an internal stem with radial fins. The effect of change in the geometric distribution of the TCEs on the performance of the heat sink was studied while maintaining the total volume of TCEs constant. The ratio of the volume of the TCEs to the cavity was also fixed constant at 10%. Initially, experiments were carried out on a cylindrical heat sink with a PCM fill ratio of 99%. A constant heat input of 6 W was applied at the bottom of the heat sink through an electrical heater. The PCM used is n-eicosane and the heat sink is made of aluminium. Numerically obtained results from ANSYS Fluent 15.0 were compared with in-house experimental results to determine the heat transfer coefficient from the walls of the heat sink. The heat sink and TCE parameters namely the diameter of the cavity, the height of the cavity, the diameter of the stem, the number of fins and the thickness of fins were treated as design variables. For a constant power level, volume of cavity and volume of TCEs, 25 different geometric configurations (initial population) were generated and the charging and discharging times were obtained as outputs by full three dimensional transient numerical simulations. Artificial Neural Networks (ANNs) were then used to create a fast forward model and multi objective optimization was carried out to obtain the optimal set of geometric configurations which simultaneously maximizes the charging time and minimize the discharging time. The optimal solutions thus obtained were validated with full numerical simulations and were also compared with the initial population. Additional insights into the mechanism of heat transfer and fluid flow were obtained for the problem under consideration. © 2017