Persistent Search: A Bridge between Depth-First and Breadth-First Search for Physical Agents.

摘要

Current search algorithms and heuristics perform very poorly in the highly realistic scenario of a physical agent traversing an initially unknown search space. They do not attempt to minimize the amount of movement required by the physical agent attempting to reach a desired goal location. In order to overcome the failings of these algorithms in dealing with searches of this particular nature, a new algorithm called persistent search was created. Persistent search differs from most other algorithms because it focuses on minimizing the physical movement of an active agent traversing an unknown search space, coping with the physical aspects of the problem which are too often ignored. Persistent search uses several standard search techniques but applies them in such a way as to change the semantics of the search. An interesting additional property of this algorithm is that through the manipulation of a single control variable, termed the persistence factor, the operation of the basic algorithm can be changed to span the continuum of behaviors between depth-first and breadth-first search. Keywords: Artificial intelligence; Search; Path planning; Complexity analysis; Target tracking. (kt)