Fast-Forward Heuristic for Multiagent Planning

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21230%2F13%3A00206027" target="_blank" >RIV/68407700:21230/13:00206027 - isvavai.cz</a>

Výsledek na webu

<a href="http://icaps13.icaps-conference.org/wp-content/uploads/2013/05/dmap13-proceedings.pdf" target="_blank" >http://icaps13.icaps-conference.org/wp-content/uploads/2013/05/dmap13-proceedings.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Fast-Forward Heuristic for Multiagent Planning

Popis výsledku v původním jazyce

Use of heuristics in search-based domain-independent deterministic multiagent planning is as important as in classical planning. In this work we propose a formal and an algorithmic adaptation of a well-known heuristic Fast-Forward into multiagent planning. Such treatment is important as it solves challenges in decentralization of this and other heuristics based on relaxation of the original planning problem. Such decentralization enables global heuristic estimates to be computed without exposing local information. Additionally, since Fast-Forward heuristic is based on relaxed planning, we propose a multiagent approach for building factored relaxed planning graphs among the agents. We sketch proofs that the results of the distributed version of the algorithm gives the same results as the centralized version. Finally, we experimentally validate dierent distribution strategies of the heuristic estimate.

Název v anglickém jazyce

Fast-Forward Heuristic for Multiagent Planning

Popis výsledku anglicky

Use of heuristics in search-based domain-independent deterministic multiagent planning is as important as in classical planning. In this work we propose a formal and an algorithmic adaptation of a well-known heuristic Fast-Forward into multiagent planning. Such treatment is important as it solves challenges in decentralization of this and other heuristics based on relaxation of the original planning problem. Such decentralization enables global heuristic estimates to be computed without exposing local information. Additionally, since Fast-Forward heuristic is based on relaxed planning, we propose a multiagent approach for building factored relaxed planning graphs among the agents. We sketch proofs that the results of the distributed version of the algorithm gives the same results as the centralized version. Finally, we experimentally validate dierent distribution strategies of the heuristic estimate.

Druh

O - Ostatní výsledky

CEP obor

JC - Počítačový hardware a software

OECD FORD obor

—

Projekt

—

Návaznosti

V - Vyzkumna aktivita podporovana z jinych verejnych zdroju

Rok uplatnění

2013

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů