Combining Search and Analogical Reasoning in Path Planning from Road Maps

摘要: Path planning from road maps is a task that may involve multiple goal interactions and ways of achieving goal. This problem recognized as difficult solving task. In this domain it particularly interesting to explore learning techniques can improve the solver’s efficiency both at plan generation execution. We want study two particular novel angles: real execution in an autonomous vehicle (instead simulated execution); interspersing replanning additional experience. paper presents initial work towards goal, namely integration analogical reasoning with when applied path large maps. show how complexity related goals. review case representation describe these cases are reused where we interleave breadth-first search technique replay. Finally, empirical results using map.