サッカーエージェントによる協調パスプレイの生成

• 秋山, 直之

• 鈴木, 恵二

• 山本, 雅人

• 大内, 東

書誌事項

タイトル別名

• Creating Cooperative Pass Play with Soccer Simulation Agent

抄録

The purpose of this research is to create the cooperative pass play with soccer simulation agent. The cooperative pass play, like "wall pass" or "through pass", needs to synchronize passer's intending space to pass with receiver's intending space to receive the pass. In this research, we propose the method for realizing the play. In addition, we adopt the hybrid learning system-combined neural network, learning classifier system, and Q-learning-proposed by Robert E. Smith in order that soccer agents may learn autonomously to decide appropriate space in accordance with the environmental states to complete pass play. By using the agent implemented the method, we try to realize the cooperative pass play in the two-on-one situation on the RoboCup Soccer Server.

The purpose of this research is to create the cooperative pass play with soccer simulation agent. The cooperative pass play, like "wall pass" or "through pass", needs to synchronize passer's intending space to pass with receiver's intending space to receive the pass. In this research, we propose the method for realizing the play. In addition, we adopt the hybrid learning system-combined neural network, learning classifier system, and Q-learning-proposed by Robert E. Smith in order that soccer agents may learn autonomously to decide appropriate space in accordance with the environmental states to complete pass play. By using the agent implemented the method, we try to realize the cooperative pass play in the two-on-one situation on the RoboCup Soccer Server.

収録刊行物

• ゲームプログラミングワークショップ1999論文集

  ゲームプログラミングワークショップ1999論文集 1999 (14), 193-200, 1999-10-15

詳細情報

• CRID: 1050292572137409280
• NII論文ID: 170000080118
• Web Site: http://id.nii.ac.jp/1001/00097451/
• 本文言語コード: ja
• 資料種別: conference paper
• データソース種別:
  • IRDB
  • CiNii Articles