Xiaoxun Sun

Comparing real-time and incremental heuristic search for real-time situated agents

Real-time situated agents, such as characters in real-time computer games, often do not know the terrain in advance but automatically observe it within a certain range around themselves. They have to interleave searches with action executions to make the searches tractable when moving autonomously to user-specified coordinates. The searches face real-time requirements since it is important that the agents be responsive to the commands of the users and move smoothly. In this article, we compare two classes of fast heuristic search methods for these navigation tasks that speed up A* searches in different ways, namely real-time heuristic search and incremental heuristic search, to understand their advantages and disadvantages and make recommendations about when each one should be used. We first develop a competitive real-time heuristic search method. LSS-LRTA* is a version of Learning Real-Time A* that uses A* to determine its local search spaces and learns quickly. We analyze the properties of LSS-LRTA* and then compare it experimentally against the state-of-the-art incremental heuristic search method D* Lite on our navigation tasks, for which D* Lite was specifically developed, resulting in the first comparison of real-time and incremental heuristic search in the literature. We characterize when to choose each one of the two heuristic search methods, depending on the search objective and the kind of terrain. Our experimental results show that LSS-LRTA* can outperform D* Lite under the right conditions, namely when there is time pressure or the user-supplied h-values are generally not misleading.

Speeding up moving-target search

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.

Incremental DCOP Search Algorithms for Solving Dynamic DCOP Problems

Distributed constraint optimization (DCOP) problems are well-suited for modeling multi-agent coordination problems. However, it only models static problems, which do not change over time. Consequently, researchers have introduced the Dynamic DCOP (DDCOP) model to model dynamic problems. In this paper, we make two key contributions: (a) a procedure to reason with the incremental changes in DDCOP problems and (b) an incremental pseudo-tree construction algorithm that can be used by DCOP algorithms such as any-space ADOPT and any-space BnB-ADOPT to solve DDCOP problems. Due to the incremental reasoning employed, our experimental results show that any-space ADOPT and any-space BnB-ADOPT are up to 42% and 38% faster, respectively, with the incremental procedure and the incremental pseudo-tree reconstruction algorithm than without them.

Reusing cost-minimal paths for goal-directed navigation in partially known terrains

Situated agents frequently need to solve search problems in partially known terrains in which the costs of the arcs of the search graphs can increase (but not decrease) when the agents observe new information. An example of such search problems is goal-directed navigation with the freespace assumption in partially known terrains, where agents repeatedly follow cost-minimal paths from their current locations to given goal locations. Incremental heuristic search is an approach for solving the resulting sequences of similar search problems potentially faster than with classical heuristic search, by reusing information from previous searches to speed up its current search. There are two classes of incremental heuristic search algorithms, namely those that make the

Moving target D* Lite

h

The fringe-saving A* search algorithm: a feasibility study

h-values of the current search more informed (such as Adaptive A*) and those that reuse parts of the A* search trees of previous searches during the current search (such as D* Lite). In this article, we introduce Path-Adaptive A* and its generalization Tree-Adaptive A*. Both incremental heuristic search algorithms terminate their searches before they expand the goal state, namely when they expand a state that is on a provably cost-minimal path to the goal. Path-Adaptive A* stores a single cost-minimal path to the goal state (the reusable path), while Tree-Adaptive A* stores a set of cost-minimal paths to the goal state (the reusable tree), and is thus potentially more efficient than Path-Adaptive A* since it uses information from all previous searches and not just the last one. Tree-Adaptive A* is the first incremental heuristic search algorithm that combines the principles of both classes of incremental heuristic search algorithms. We demonstrate experimentally that both Path-Adaptive A* and Tree-Adaptive A* can be faster than Adaptive A* and D* Lite, two state-of-the-art incremental heuristic search algorithms for goal-directed navigation with the freespace assumption.