The power unit design was completed utilizing the multi objective optimization based on a neighborhood cultivation genetic algorithm. The pressure drop, thermal resistance, strain, fin height and flow velocity were examined simultaneously as the design objectives, it was able to visualize these correlations and trade-off relations. The design proposal could be chosen from the Pareto solution using the weight computed by a paired comparison of the objective functions. This approach could visualize a design process and improve the quality and speed of a design, compared with the conventional series design which determines each evaluate function in order.
View full abstract