Separate Generation and Evaluation for Parallel Greedy Best-First Search

2024-08-11

arXiv.org

AI Summary

To address thread idleness and trajectory divergence in parallel Greedy Best-First Search (GBFS) caused by the Best-First Transition Synchronization (BTS) constraint, this paper proposes the first generate–evaluate decoupled architecture. It separates state generation from heuristic evaluation, enabling asynchronous state generation and priority-driven distributed evaluation—thereby eliminating BTS synchronization overhead without compromising search equivalence. Grounded in Bench Transition System theory, the approach formally guarantees state coverage consistency with sequential GBFS. Experimental evaluation across multiple classical heuristic search benchmarks demonstrates a 2.3–4.1× improvement in state evaluation throughput and an average 37% reduction in search time.

Technology Category

• Search and Optimization: Heuristic Search
• Constraint Satisfaction and Optimization: Search
• Multiagent Systems: Distributed Problem Solving

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Parallelization of Greedy Best First Search (GBFS) has been difficult because straightforward parallelization can result in search behavior which differs significantly from sequential GBFS, exploring states which would not be explored by sequential GBFS with any tie-breaking strategy. Recent work has proposed a class of parallel GBFS algorithms which constrains search to exploration of the Bench Transition System (BTS), which is the set of states that can be expanded by GBFS under some tie-breaking policy. However, enforcing this constraint is costly, as such BTS-constrained algorithms are forced to spend much of the time waiting so that only states which are guaranteed to be in the BTS are expanded. We propose an improvement to parallel search which decouples state generation and state evaluation and significantly improves state evaluation rate, resulting in better search performance.

Problem

• Improve constrained parallel greedy best-first search
• Decouple state generation and evaluation
• Enhance search performance via faster evaluation

Innovation

• Decouples state generation and evaluation
• Improves state evaluation rate
• Enhances parallel search performance