بهینه سازی موازی کلونی مورچه ها برای تعیین موقعیت منبع حرارت نقطه در یک دامنه 2-D

The Ant Colony Optimization (ACO) and its parallel implementations are applied to determine the unknown position of a point heat source in a two-dimensional steady-state heat conduction problem. The heuristic value, the determination of path and the objective function are all studied based on the inverse heat source problem. Some of the standard steps of the ACO are also modified according to the features of the heat conduction problem. In order to accelerate the speed of solving the inverse problem, two kinds of parallel ACO strategies, the fine-grained master-slave strategy and the coarse-grained strategy combined with the above modification are studied in this paper. The results show that the fine-grained strategy has a higher speedup than the coarse-grained strategy whiles the coarse-grained strategy has a higher accuracy rate. This means that the coarse-grained strategy is more proper during the parallel implementation for solving the inverse heat source problem and the parallel implementation of the ACO can improve the computational efficiency.