Classic best-first heuristic search algorithms, like A*, record every unique state they encounter in RAM, making them infeasible for solving large problems. In this paper, we demonstrate how best-first search can be scaled to solve much larger problems by exploiting disk storage and parallel processing and, in some cases, slightly relaxing the strict best-first node expansion order. Some previous disk-based search algorithms abandon best-first search order in an attempt to increase efficiency. We present two case studies showing that A*, when augmented with Delayed Duplicate Detection, can actually be more efficient than these non-best-first search orders. First, we present a straightforward external variant of A*, called PEDAL, that slightly relaxes best-first order in order to be I/O efficient in both theory and practice, even on problems featuring real-valued node costs. Because it is easy to parallelize, PEDAL can be faster than in-memory IDA* even on domains with few duplicate states, such as the sliding-tile puzzle. Second, we present a variant of PEDAL, called PE2A*, that uses partial expansion to handle problems that have large branching factors. When tested on the problem of Multiple Sequence Alignment, PE2A* is the first algorithm capable of solving the entire Reference Set 1 of the standard BAliBASE benchmark using a biologically accurate cost function. This work shows that classic best-first algorithms like A* can be applied to large real-world problems. We also provide a detailed implementation guide with source code both for generic parallel disk-based best-first search and for Multiple Sequence Alignment with a biologically accurate cost function. Given its effectiveness as a general-purpose problem-solving method, we hope that this makes parallel and disk-based search accessible to a wider audience.

In real-time domains such as video games, planning happens concurrently with execution and the planning algorithm has a strictly bounded amount of time before it must return the next action for the agent to execute. We explore the use of real-time heuristic search in two benchmark domains inspired by video games. Unlike classic benchmarks such as grid pathfinding and the sliding tile puzzle, these new domains feature exogenous change and directed state space graphs. We consider the setting in which planning and acting are concurrent and we use the natural objective of minimizing goal achievement time. Using both the classic benchmarks and the new domains, we investigate several enhancements to a leading real-time search algorithm, LSS-LRTA*. We show experimentally that 1) it is better to plan after each action or to use a dynamically sized lookahead, 2) A*-based lookahead can cause undesirable actions to be selected, and 3) on-line de-biasing of the heuristic can lead to improved performance. We hope this work encourages future research on applying real-time search in dynamic domains.

In real-time domains such as video games, a planning algo- rithm has a strictly bounded time before it must return the next action for the agent to execute. We introduce a realistic video game benchmark domain that is useful for evaluating real-time heuristic search algorithms. Unlike previous bench- marks such as grid pathfinding and the sliding tile puzzle, this new domain includes dynamics and induces a directed graph. Using both the previous and new domains, we investigate sev- eral enhancements to a leading real-time search algorithm, LSS-LRTA*. We show experimentally that 1) it is not dif- ficult to outperform A* when optimizing goal achievement time, 2) it is better to plan after each action than to commit to multiple actions or to use a dynamically sized lookahead, 3) A*-based lookahead can cause undesirable actions to be selected, and 4) on-line de-biasing of the heuristic can lead to improved performance. We hope that this new domain and results will stimulate further research on applying real-time search to dynamic real-time domains.

In many applications of shortest-path algorithms, it is impractical to find a provably optimal solution; one can only hope to achieve an appropriate balance between search time and solution cost that respects the user's preferences. Preferences come in many forms; we consider utility functions that linearly trade-off search time and solution cost. Many natural utility functions can be expressed in this form. For example, when solution cost represents the makespan of a plan, equally weighting search time and plan makespan minimizes the time from the arrival of a goal until it is achieved. Current state-of-the-art approaches to optimizing utility functions rely on anytime algorithms, and the use of extensive training data to compute a termination policy. We propose a more direct approach, called Bugsy, that incorporates the utility function directly into the search, obviating the need for a separate termination policy. We describe a new method based on off-line parameter tuning and a novel benchmark domain for planning under time pressure based on platform-style video games. We then present what we believe to be the first empirical study of applying anytime monitoring to heuristic search, and we compare it with our proposals. Our results suggest that the parameter tuning technique can give the best performance if a representative set of training instances is available. If not, then Bugsy is the algorithm of choice, as it performs well and does not require any off-line training. This work extends the tradition of research on metareasoning for search by illustrating the benefits of embedding lightweight reasoning about time into the search algorithm itself.

Heuristic search is a technique used pervasively in the fieldsof artificial intelligence, automated planning and operations research to solve a wide range of problems from planning military deployments to planning tasks for a robot that must clean a messy kitchen. An automated agent can use heuristic search to construct a plan that, when executed, will achieve a desired task. The search algorithm explores different sequences of actions that the agent can execute, looking for a sequence that will lead it to a desired goal state. In many situations, an agent is given a task that it would like to solve as quickly as possible. The agent must allocate its time between searching for the actions that will achieve the task and actually executing them. We call this problem planning under time pressure.

Published papers rarely disclose implementation details. In this paper we show how such details can account for speedups of up to a factor of 28 for different implementations of the same algorithm. We perform an in-depth analysis of the most popular benchmark in heuristic search: the 15-puzzle. We study implementation choices in C++ for both IDA* and A* using the Manhattan distance heuristic. Results suggest that several optimizations deemed critical in folklore provide only small improvements while seemingly innocuous choices can play a large role. These results are important for ensuring that the correct conclusions are drawn from empirical comparisons

Motion planning in continuous space is a fundamentalrobotics problem that has been approached from many per-spectives. Rapidly-exploring Random Trees (RRTs) usesampling to efficiently traverse the continuous and high-dimensional state space. Heuristic graph search methods uselower bounds on solution cost to focus effort on portions ofthe space that are likely to be traversed by low-cost solutions.In this work, we bring these two ideas together in a tech-nique called f -biasing: we use estimates of solution cost,computed as in heuristic search, to guide sparse sampling,as in RRTs. We see this new technique as strengthening theconnections between motion planning in robotics and combi-natorial search in artificial intelligence.

The memory requirements of basic best-first heuristic search algorithms like A* make them infeasible for solving large problems. External disk storage is cheap and plentiful com- pared to the cost of internal RAM. Unfortunately, state-of- the-art external memory search algorithms either rely on brute-force search techniques, such as breadth-first search, or they rely on all node values falling in a narrow range of in- tegers, and thus perform poorly on real-world domains with real-valued costs. We present a new general-purpose algo- rithm, PEDAL, that uses external memory and parallelism to perform a best-first heuristic search capable of solving large problems with real costs. We show theoretically that PEDAL is I/O efficient and empirically that it is both better on a stan- dard unit-cost benchmark, surpassing internal IDA* on the 15-puzzle, and gives far superior performance on problems with real costs.

Tree search is a common technique for solving constraint satisfaction and combinatorial optimization problems. The most popular strategies are depth-first search and limited discrepancy search. Aside from pruning or ordering the children of each node, these algorithms do not adapt their search order to take advantage of information that becomes available during search, such as heuristic scores or leaf costs. We present a framework called best-leaf-first search (BLFS) that uses this additional information to estimate the cost of taking discrepancies in the search tree and then attempts to visit leaves in a best-first order. In this way, BLFS brings the idea of best-first search from shortest path problems to the areas of constraint satisfaction and combinatorial optimization. Empirical results demonstrate that this new dynamic approach results in better search performance than previous static search strategies on two very different domains: structured CSPs and the traveling salesman problem.

In problem domains where an informative heuristic evaluation function is not known or not easily computed, abstraction can be used to derive admissible heuristic values. Optimal path lengths in the abstracted problem are consistent heuristic estimates for the original problem. Pattern databases are the traditional method of creating such heuristics, but they exhaustively compute costs for all abstract states and are thus usually appropriate only when all instances share the same single goal state. Hierarchical heuristic search algorithms address these shortcomings by searching for paths in the abstract space on an as-needed basis. However, existing hierarchical algorithms search less efficiently than pattern database constructors: abstract nodes may be expanded many times during the course of a base-level search. We present a novel hierarchical heuristic search algorithm, called Switchback, that uses an alternating direction of search to avoid abstract node re-expansions. This algorithm is simple to implement and demonstrates superior performance to existing hierarchical heuristic search algorithms on several standard benchmarks.

To harness modern multicore processors, it is imperative to develop parallel versions of fundamental algorithms. In this paper, we compare different approaches to parallel best-first search in a shared-memory setting. We present a new method, PBNF, that uses abstraction to partition the state space and to detect duplicate states without requiring frequent locking. PBNF allows speculative expansions when necessary to keep threads busy. We identify and fix potential livelock conditions in our approach, proving its correctness using temporal logic. Our approach is general, allowing it to extend easily to suboptimal and anytime heuristic search. In an empirical comparison on STRIPS planning, grid pathfinding, and sliding tile puzzle problems using 8-core machines, we show that A*, weighted A* and Anytime weighted A* implemented using PBNF yield faster search than improved versions of previous parallel search proposals.