Garrett M. Fereday, Eric A. Hansen

A Bucket-Based Priority Queue for Bounded-Suboptimal and Anytime A* Search

Introduction

A bucket-based priority queue for bounded-suboptimal and anytime a* search. Introduces a bucket heap, a new priority queue generalizing the bucket queue to accelerate bounded-suboptimal and Anytime A* search algorithms for shortest-path problems.

Abstract

We introduce a priority queue data structure, called a bucket heap, which generalizes the bucket queue commonly used to accelerate A* search for shortest-path problems with a small range of integer transition costs. Unlike a bucket queue, a bucket heap speeds up priority queue operations for bounded-suboptimal and Anytime A* algorithms guided by non-admissible node evaluation functions. It also provides direct access---without any additional overhead---to the underlying bucket queue of A*, which we show can be used to improve search performance in further ways.

Review

The paper introduces a novel data structure, the "bucket heap," designed to significantly enhance the performance of A* search algorithms. Positioned as a generalization of the widely-used bucket queue, which effectively accelerates A* for shortest-path problems with small integer transition costs, this work extends such efficiency benefits to a broader and more complex class of search scenarios. The core contribution lies in adapting the bucket queue concept to address computational bottlenecks in modern, practical A* variants. A key distinction and advantage of the bucket heap lies in its specific optimization for bounded-suboptimal and Anytime A* algorithms. Unlike traditional bucket queues, this new structure is engineered to work effectively with non-admissible node evaluation functions, which are often employed in these advanced A* variants to prioritize speed over strict optimality. Furthermore, the abstract highlights an intriguing design feature: the bucket heap provides direct, overhead-free access to its underlying bucket queue, a capability that is claimed to unlock additional avenues for improving search performance beyond the direct acceleration of priority queue operations. Overall, the proposed bucket heap presents a promising development for the field of heuristic search. By specifically targeting the computational needs of bounded-suboptimal and Anytime A* search, this data structure could lead to significant practical speedups in applications where solutions within certain optimality bounds or available time limits are paramount. The claims of direct access to the underlying bucket queue for further performance enhancements hint at a well-thought-out and potentially versatile design, positioning this as a noteworthy contribution to the optimization of advanced A* search strategies.

Full Text

Comments

You need to be logged in to post a comment.