Adaptive A* (Koenig and Likhachev 2005) Setting: repeatedly find shortest path to a fixed set of goal states, but varying start states. Intuitively, it uses that after making a search from \(s\), we know that all states close to \(s\) must have a distance that is not much smaller than the distance from \(s\). Input: a consistent heuristic \(H(s, t)\) for any pair of states, that additionally satisfies the more general triangle inequality.