Speeding up moving-target search

Abstract

In this paper, we study moving-target search, where an agent (= hunter) has to catch a moving target (= prey). The agent does not necessarily know the terrain initially but can observe it within a certain sensor range around itself. It uses the strategy to always move on a shortest presumed unblocked path toward the target, which is a reasonable strategy for computer-controlled characters in video games. We study how the agent can find such paths faster by exploiting the fact that it performs A* searches repeatedly. To this end, we extend Adaptive A*, an incremental heuristic search method, to moving-target search and demonstrate experimentally that the resulting MT-Adaptive A* is faster than isolated A* searches and, in many situations, also D* Lite, a state-of-the-art incremental heuristic search method. In particular, it is faster than D* Lite by about one order of magnitude for moving-target search in known and initially unknown mazes if both search methods use the same informed heuristics.

Planning When Goals Change: A Moving Target Search Approach

“…This approach was shown to be complete in turns based settings when the target periodically skips moves but it is subject to heuristic depressions and lost of information when the target moves [7]. Currently, the state-of-the-art algorithms with this first strategy are variants of the AA* algorithm [8]: MTAA* [9] and GAA* [10]. All these algorithms must use admissible heuristics to ensure their soundness and completeness.…”

Hierarchical visibility for guaranteed search in large-scale outdoor terrain

“…The assumptions about target motion differs dramatically and targets generally move at bounded speed and their location is known by the searcher. In [27] an A* planner with the graph representing a grid with obstacles is used to solve the problem. Moldenhauer et al [36] presented the Dynamic Abstract TrailMax algorithm for computing strategies in moving target search based on PartialRefinement A* (PRA*) planning [44].…”

Multi-agent real-time pursuit

“…One solution is to make off-line algorithms to be incremental [14], which is a continual planning technique that make use of information from previous searches to find solutions to the problems potentially faster than are possible by solving the problems from scratch. D* [15,16], Focused D* [17], D* Lite [18][19][20] and MT-Adaptive A* [21] are some of the well-known optimal incremental heuristic search algorithms applied to path planning domain. These algorithms are efficient in most cases, but sometimes a small change in the environment may cause to re-plan almost a complete path from scratch, which requires polynomial time and does not meet real-time constraints.…”

“…18 for illustration). Finally we compute the utility of the second best alternative (lines [20][21][22][23][24][25][26], and if the utility is greater than or equal to 1, we select the second best direction, otherwise we select the best direction as the final proposed moving direction (lines [27][28][29][30]. The utility formula increases when attraction direction is closer to the second best direction or the estimated path length of second best route is closer to the best one; and is determined such that the second best alternative route may be selected only when it is at most two times longer than the best one.…”