English

Iterative Budgeted Exponential Search

Data Structures and Algorithms 2019-07-31 v1 Artificial Intelligence

Abstract

We tackle two long-standing problems related to re-expansions in heuristic search algorithms. For graph search, A* can require Ω(2n)\Omega(2^{n}) expansions, where nn is the number of states within the final ff bound. Existing algorithms that address this problem like B and B' improve this bound to Ω(n2)\Omega(n^2). For tree search, IDA* can also require Ω(n2)\Omega(n^2) expansions. We describe a new algorithmic framework that iteratively controls an expansion budget and solution cost limit, giving rise to new graph and tree search algorithms for which the number of expansions is O(nlogC)O(n \log C), where CC is the optimal solution cost. Our experiments show that the new algorithms are robust in scenarios where existing algorithms fail. In the case of tree search, our new algorithms have no overhead over IDA* in scenarios to which IDA* is well suited and can therefore be recommended as a general replacement for IDA*.

Keywords

Cite

@article{arxiv.1907.13062,
  title  = {Iterative Budgeted Exponential Search},
  author = {Malte Helmert and Tor Lattimore and Levi H. S. Lelis and Laurent Orseau and Nathan R. Sturtevant},
  journal= {arXiv preprint arXiv:1907.13062},
  year   = {2019}
}
R2 v1 2026-06-23T10:35:05.974Z