In this paper, multi objective Pareto optimization of frost formation in Micro Channel Heat Sink (MCHS) with circular fins with square and triangular configurations is performed in three steps. In the first step, the fluid flow is solved numerically in a number of MCHS using CFD techniques. In numerical modeling, basic equations of humid air and frost, including continuum, momentum, energy and phase change mechanism are numerically solved. Flow type is transient and the microfluidic regime is slip flow. In the second step, CFD data is used to model the objective functions namely volume fraction of ice and the heat transfer in MCHS using GMDH type artificial neural networks. The accuracy of GMDH modeling of objective functions is investigated using different statistical models. Finally, polynomials extracted from the GMDH are used to multi-objective optimization of frost formation in MCHS using NSGA II algorithm. In this step, the Pareto front, which contains very useful information in the design of MCHS containing frost formation, will be extracted and deeply discussed. In the results section, points with unique features are introduced. It will be shown that the design information presented in the results about MCHS containing frost formation can only be extracted with a multi-objective optimization approach.
