In this work, the authors propose an interesting approach called lifelong planning A* (LPA*). It consists of combining an incremental version of the well-known A* with other artificial intelligence (AI) techniques (namely, dynamic strict weakly superior function - fixed point (DynamicSWSF-FP)) to quickly find the shortest paths between a pair of vertices of a given graph (namely, path planning problems).

Many experiments are conducted to prove the efficiency of this approach. The argument given by the authors to show that their technique is faster than the classical ones is correct. In fact, their approach didn’t start from scratch each time during the resolution process to find the other shortest paths. Previous search results are simply reused, to reach the target vertex differently. Thus, the proposed technique is efficient in large domains with frequent changes. The paper offers a good overview of the path planning problem; it is also easy to read and understand.