When to Commit to an Action in Online Planning and Search.

Tianyi Gu; Wheeler Ruml; Shahaf S. Shperberg; Eyal Solomon Shimony; Erez Karpas

doi:10.1609/socs.v15i1.21755

When to Commit to an Action in Online Planning and Search.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

In online planning, search is concurrent with execution. Un-der the formulation of planning as heuristic search, when a planner commits to an action, it re-roots its search tree at the node representing the outcome of that action. For the system to remain controlled, the planner must commit to a new action (possibly a no-op) before the previously chosen action completes. This time pressure results in a real-time search. In this time-bounded setting, it can be beneficial to commit early, in order to perform more lookahead search focused be-low an upcoming state. In this paper, we propose a principled method for making this commitment decision. Our experimental evaluation shows that our scheme can outperform previously-proposed fixed strategies.

Original language	English
Title of host publication	SoCS
Subtitle of host publication	Proceedings of the International Symposium on Combinatorial Search
Pages	83-90
Number of pages	8
Volume	15
Edition	1
DOIs	https://doi.org/10.1609/socs.v15i1.21755
State	Published - 17 Jul 2022

Keywords

Combinatorial Puzzles
Meta Reasoning And Search
Problem Solving Using Search
Real-time Search

Access to Document

10.1609/socs.v15i1.21755

Cite this

@inproceedings{e88f1d18b63e40b49da5c0c0015e9f52,

title = "When to Commit to an Action in Online Planning and Search.",

abstract = "In online planning, search is concurrent with execution. Un-der the formulation of planning as heuristic search, when a planner commits to an action, it re-roots its search tree at the node representing the outcome of that action. For the system to remain controlled, the planner must commit to a new action (possibly a no-op) before the previously chosen action completes. This time pressure results in a real-time search. In this time-bounded setting, it can be beneficial to commit early, in order to perform more lookahead search focused be-low an upcoming state. In this paper, we propose a principled method for making this commitment decision. Our experimental evaluation shows that our scheme can outperform previously-proposed fixed strategies.",

keywords = "Combinatorial Puzzles, Meta Reasoning And Search, Problem Solving Using Search, Real-time Search",

author = "Tianyi Gu and Wheeler Ruml and Shperberg, \{Shahaf S.\} and Shimony, \{Eyal Solomon\} and Erez Karpas",

note = "DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.",

year = "2022",

month = jul,

day = "17",

doi = "10.1609/socs.v15i1.21755",

language = "English",

isbn = "1577358732",

volume = "15",

pages = "83--90",

booktitle = "SoCS",

edition = "1",

}

TY - GEN

T1 - When to Commit to an Action in Online Planning and Search.

AU - Gu, Tianyi

AU - Ruml, Wheeler

AU - Shperberg, Shahaf S.

AU - Shimony, Eyal Solomon

AU - Karpas, Erez

N1 - DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.

PY - 2022/7/17

Y1 - 2022/7/17

N2 - In online planning, search is concurrent with execution. Un-der the formulation of planning as heuristic search, when a planner commits to an action, it re-roots its search tree at the node representing the outcome of that action. For the system to remain controlled, the planner must commit to a new action (possibly a no-op) before the previously chosen action completes. This time pressure results in a real-time search. In this time-bounded setting, it can be beneficial to commit early, in order to perform more lookahead search focused be-low an upcoming state. In this paper, we propose a principled method for making this commitment decision. Our experimental evaluation shows that our scheme can outperform previously-proposed fixed strategies.

AB - In online planning, search is concurrent with execution. Un-der the formulation of planning as heuristic search, when a planner commits to an action, it re-roots its search tree at the node representing the outcome of that action. For the system to remain controlled, the planner must commit to a new action (possibly a no-op) before the previously chosen action completes. This time pressure results in a real-time search. In this time-bounded setting, it can be beneficial to commit early, in order to perform more lookahead search focused be-low an upcoming state. In this paper, we propose a principled method for making this commitment decision. Our experimental evaluation shows that our scheme can outperform previously-proposed fixed strategies.

KW - Combinatorial Puzzles

KW - Meta Reasoning And Search

KW - Problem Solving Using Search

KW - Real-time Search

U2 - 10.1609/socs.v15i1.21755

DO - 10.1609/socs.v15i1.21755

M3 - Conference contribution

SN - 1577358732

VL - 15

SP - 83

EP - 90

BT - SoCS

ER -

When to Commit to an Action in Online Planning and Search.

Abstract

Keywords

Access to Document

Fingerprint

Cite this