Accelerating A * algorithms by sweeping out small-degree nodes

A* is an algorithm framework for calculating point-to-point shortest paths. This paper gives a simple method to accelerate A* algorithms in practice by sweeping out small-degree nodes from the priority queue, which can reduce the running time of the queue operations and the distance estimations. Experiments show that our method is efficient in practice, especially for A* algorithms with a heavy estimation function such as the ALT algorithm (Goldberg and Harrelson, SODA 2005) and its time-dependent generalizations.