人工知能学会第二種研究会資料
Online ISSN : 2436-5556
RRS における L-DCOP モデルを用いた異種エージェント間協調について
鈴木 宏哉上原 温揮藤澤 丈前田 綾也松永 一希安藤 圭祐内種 岳詞岩田 員典伊藤 暢浩
著者情報
研究報告書・技術報告書 フリー

2024 年 2023 巻 SAI-047 号 p. 03-

詳細
抄録

RRS (RoboCupRescue Simulation) is a disaster relief simulation for urban earthquakes. The RRS has some rescue tasks and some kinds of agents to solve them. This simulation must determine a proper assignment between agents and rescue tasks to rescue civilians as much as possible. It is important for such a disaster relief problem to consider the rescue order of the tasks and time constraints of the tasks and to cooperate among some kinds of agents. This study models the disaster relief problem of RRS with order and time constraints as the Layered DCOP (L-DCOP). The L-DCOP is an extension of DCOP whose definition includes order and time constraints. We designed and implemented disaster relief agents that solve the L-DCOP in RRS. As a result of some simulations under various conditions, we confirmed that some kinds of agents could cooperate appropriately based on the order and time constraints.

著者関連情報
© 2024 著作者
次の記事
feedback
Top