Results 1  10
of
32
The generalized A* architecture
 Journal of Artificial Intelligence Research
, 2007
"... We consider the problem of computing a lightest derivation of a global structure using a set of weighted rules. A large variety of inference problems in AI can be formulated in this framework. We generalize A * search and heuristics derived from abstractions to a broad class of lightest derivation p ..."
Abstract

Cited by 47 (6 self)
 Add to MetaCart
We consider the problem of computing a lightest derivation of a global structure using a set of weighted rules. A large variety of inference problems in AI can be formulated in this framework. We generalize A * search and heuristics derived from abstractions to a broad class of lightest derivation problems. We also describe a new algorithm that searches for lightest derivations using a hierarchy of abstractions. Our generalization of A * gives a new algorithm for searching AND/OR graphs in a bottomup fashion. We discuss how the algorithms described here provide a general architecture for addressing the pipeline problem — the problem of passing information back and forth between various stages of processing in a perceptual system. We consider examples in computer vision and natural language processing. We apply the hierarchical search algorithm to the problem of estimating the boundaries of convex objects in grayscale images and compare it to other search methods. A second set of experiments demonstrate the use of a new compositional model for finding salient curves in images. 1.
Casebased subgoaling in realtime heuristic search for video game pathfinding
 J. Artif. Intell. Res
, 2010
"... Realtime heuristic search algorithms satisfy a constant bound on the amount of planning per action, independent of problem size. As a result, they scale up well as problems become larger. This property would make them well suited for video games where Artificial Intelligence controlled agents must ..."
Abstract

Cited by 17 (4 self)
 Add to MetaCart
(Show Context)
Realtime heuristic search algorithms satisfy a constant bound on the amount of planning per action, independent of problem size. As a result, they scale up well as problems become larger. This property would make them well suited for video games where Artificial Intelligence controlled agents must react quickly to user commands and to other agents ’ actions. On the downside, realtime search algorithms employ learning methods that frequently lead to poor solution quality and cause the agent to appear irrational by revisiting the same problem states repeatedly. The situation changed recently with a new algorithm, D LRTA*, which attempted to eliminate learning by automatically selecting subgoals. D LRTA * is well poised for video games, except it has a complex and memorydemanding precomputation phase during which it builds a database of subgoals. In this paper, we propose a simpler and more memoryefficient way of precomputing subgoals thereby eliminating the main obstacle to applying stateoftheart realtime search methods in video games. The new algorithm solves a number of randomly chosen problems offline, compresses the solutions into a series of subgoals and stores them in a database. When presented with a novel problem online, it queries the database for the most similar previously solved case and uses its subgoals to solve the problem. In the domain of pathfinding on four large video game maps, the new algorithm delivers solutions eight times better while using 57 times less memory and requiring 14 % less precomputation time. 1.
Dynamic control in realtime heuristic search
, 2008
"... Realtime heuristic search is a challenging type of agentcentered search because the agent’s planning time per action is bounded by a constant independent of problem size. A common problem that imposes such restrictions is pathfinding in modern computer games where a large number of units must plan ..."
Abstract

Cited by 16 (11 self)
 Add to MetaCart
(Show Context)
Realtime heuristic search is a challenging type of agentcentered search because the agent’s planning time per action is bounded by a constant independent of problem size. A common problem that imposes such restrictions is pathfinding in modern computer games where a large number of units must plan their paths simultaneously over large maps. Common search algorithms (e.g., A*, IDA*, D*, ARA*, AD*) are inherently not realtime and may lose completeness when a constant bound is imposed on peraction planning time. Realtime search algorithms retain completeness but frequently produce unacceptably suboptimal solutions. In this paper, we extend classic and modern realtime search algorithms with an automated mechanism for dynamic depth and subgoal selection. The new algorithms remain realtime and complete. On large computer game maps, they find paths within 7 % of optimal while on average expanding roughly a single state per action. This is nearly a threefold improvement in suboptimality over the existing stateoftheart algorithms and, at the same time, a 15fold improvement in the amount of planning per action.
Path Planning with Adaptive Dimensionality
"... Path planning quickly becomes computationally hard as the dimensionality of the statespace increases. In this paper, we present a planning algorithm intended to speed up path planning for highdimensional statespaces such as robotic arms. The idea behind this work is that while planning in a highd ..."
Abstract

Cited by 13 (7 self)
 Add to MetaCart
(Show Context)
Path planning quickly becomes computationally hard as the dimensionality of the statespace increases. In this paper, we present a planning algorithm intended to speed up path planning for highdimensional statespaces such as robotic arms. The idea behind this work is that while planning in a highdimensional statespace is often necessary to ensure the feasibility of the resulting path, large portions of the path have a lowerdimensional structure. Based on this observation, our algorithm iteratively constructs a statespace of an adaptive dimensionality–a statespace that is highdimensional only where the higher dimensionality is absolutely necessary for finding a feasible path. This often reduces drastically the size of the statespace, and as a result, the planning time and memory requirements. Analytically, we show that our method is complete and is guaranteed to find a solution if one exists, within a specified suboptimality bound. Experimentally, we apply the approach to 3D vehicle navigation (x, y, heading), and to a 7 DOF robotic arm on the Willow Garage’s PR2 robot. The results from our experiments suggest that our method can be substantially faster than some of the stateoftheart planning algorithms optimized for those tasks.
Using Distance Estimates in Heuristic Search: A Reevaluation
"... jtd7, ruml, jhg22 at cs.unh.edu Traditionally, heuristic search algorithms have relied on an estimate of the costtogo to speed up problemsolving. In many domains, operators have different costs and estimated costtogo is not the same as estimated searchdistancetogo. We investigate further acce ..."
Abstract

Cited by 7 (5 self)
 Add to MetaCart
jtd7, ruml, jhg22 at cs.unh.edu Traditionally, heuristic search algorithms have relied on an estimate of the costtogo to speed up problemsolving. In many domains, operators have different costs and estimated costtogo is not the same as estimated searchdistancetogo. We investigate further accelerating search by using a distancetogo function. We evaluate two previous proposals: Dynamically weighted A ∗ and A ∗ ǫ. We present a revision to dynamically weighted A ∗ which improves its performance substantially in domains where solutions are not at fixed depths. We show how to incorporate search distance to go estimates into weighted A ∗ in order to improve its performance in pathfinding problems. We present a proof showing that weighted A ∗ can ignore duplicate states, leading to large improvements in performance for pathfinding problems.
Hierarchical Path Planning for MultiSize Agents in Heterogeneous Environments
"... Abstract — Path planning is a central topic in games and other research areas, such as robotics. Despite this, very little research addresses problems involving agents with multiple sizes and terrain traversal capabilities. In this paper we present a new planner, Hierarchical Annotated A * (HAA*), a ..."
Abstract

Cited by 7 (2 self)
 Add to MetaCart
(Show Context)
Abstract — Path planning is a central topic in games and other research areas, such as robotics. Despite this, very little research addresses problems involving agents with multiple sizes and terrain traversal capabilities. In this paper we present a new planner, Hierarchical Annotated A * (HAA*), and demonstrate how a single abstract graph can be used to plan for agents with heterogeneous sizes and terrain traversal capabilities. Through theoretical analysis and experimental evaluation we show that HAA * is able to generate nearoptimal solutions to a wide range of problems while maintaining an exponential reduction in effort over lowlevel search. HAA * is also shown to require just a fraction of the storage space needed by the original grid map. I.
MultiDomain Realtime Planning in Dynamic Environments
"... Figure 1: Two agents navigating with spacetime precision through a complex dynamic environment. This paper presents a realtime planning framework for multicharacter navigation that enables the use of multiple heterogeneous problem domains of differing complexities for navigation in large, complex, ..."
Abstract

Cited by 5 (4 self)
 Add to MetaCart
Figure 1: Two agents navigating with spacetime precision through a complex dynamic environment. This paper presents a realtime planning framework for multicharacter navigation that enables the use of multiple heterogeneous problem domains of differing complexities for navigation in large, complex, dynamic virtual environments. The original navigation problem is decomposed into a set of smaller problems that are distributed across planning tasks working in these different domains. An anytime dynamic planner is used to efficiently compute and repair plans for each of these tasks, while using plans in one domain to focus and accelerate searches in more complex domains. We demonstrate the benefits of our framework by solving many challenging multiagent scenarios in complex dynamic environments requiring spacetime precision and explicit coordination between interacting agents, by accounting for dynamic information at all stages of the decisionmaking process.
Moving Target D * Lite ∗
"... Incremental search algorithms, such as Generalized FringeRetrieving A * and D * Lite, reuse search trees from previous searches to speed up the current search and thus often find costminimal paths for series of similar search problems faster than by solving each search problem from scratch. Howeve ..."
Abstract

Cited by 5 (2 self)
 Add to MetaCart
(Show Context)
Incremental search algorithms, such as Generalized FringeRetrieving A * and D * Lite, reuse search trees from previous searches to speed up the current search and thus often find costminimal paths for series of similar search problems faster than by solving each search problem from scratch. However, existing incremental search algorithms have limitations. For example, D * Lite is slow on moving target search problems, where both the start and goal states can change over time. In this paper, we therefore introduce Moving Target D * Lite, an extension of D * Lite that uses the principle behind Generalized FringeRetrieving A * to repeatedly calculate a costminimal path from the hunter to the target in environments whose blockages can change over time. We demonstrate experimentally that Moving Target D * Lite is four to five times faster than Generalized Adaptive A*, which so far was believed to be the fastest incremental search algorithm for solving moving target search problems in dynamic environments.
Incremental ARA*: An Incremental Anytime Search Algorithm for Moving Target Search
 In Proc. ICAPS
, 2012
"... Movingtarget search, where a hunter has to catch a moving target, is an important problem for video game developers. In our case, the hunter repeatedly moves towards the target and thus has to solve similar search problems repeatedly. We develop Incremental ARA * (IARA*) for this purpose, the f ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
(Show Context)
Movingtarget search, where a hunter has to catch a moving target, is an important problem for video game developers. In our case, the hunter repeatedly moves towards the target and thus has to solve similar search problems repeatedly. We develop Incremental ARA * (IARA*) for this purpose, the first incremental anytime search algorithm for movingtarget search in known terrain. We provide an error bound on the lengths of the paths found by IARA * and show experimentally in known fourneighbor gridworlds that IARA* can be used with smaller time limits between moves of the hunter than competing stateoftheart movingtarget search algorithms, namely repeated A*, GFRA*, FRA*, and sometimes repeated ARA*. The hunter tends to make more moves
MAPP: a Scalable MultiAgent Path Planning Algorithm with Tractability and Completeness Guarantees
"... Multiagent path planning is a challenging problem with numerous reallife applications. Running a centralized search such as A * in the combined state space of all units is complete and costoptimal, but scales poorly, as the state space size is exponential in the number of mobile units. Traditiona ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
Multiagent path planning is a challenging problem with numerous reallife applications. Running a centralized search such as A * in the combined state space of all units is complete and costoptimal, but scales poorly, as the state space size is exponential in the number of mobile units. Traditional decentralized approaches, such as FAR and WHCA*, are faster and more scalable, being based on problem decomposition. However, such methods are incomplete and provide no guarantees with respect to the running time or the solution quality. They are not necessarily able to tell in a reasonable time whether they would succeed in finding a solution to a given instance. We introduce MAPP, a tractable algorithm for multiagent path planning on undirected graphs. We present a basic version and several extensions. They have lowpolynomial worstcase upper bounds for the running time, the memory requirements, and the length of solutions. Even though all algorithmic versions are incomplete in the general case, each provides formal guarantees on problems it can solve. For each version, we discuss the algorithm’s completeness with respect to clearly defined subclasses of instances. Experiments were run on realistic game grid maps. MAPP solved 99.86 % of all mobile units, which is 18–22 % better than the percentage of FAR and WHCA*. MAPP marked 98.82 % of all units as provably solvable during the first stage of plan computation. Parts of MAPP’s computation can be reused across instances on the same map. Speedwise, MAPP is competitive or significantly faster than WHCA*, depending on whether MAPP performs all computations from scratch. When data that MAPP can reuse are preprocessed offline and readily available, MAPP is slower than the very fast FAR algorithm by a factor of 2.18 on average. MAPP’s solutions are on average 20% longer than FAR’s solutions and 7–31 % longer than WHCA*’s solutions. 1.