強化学習に基づく人の流れのシミュレーションのためのエージェントモデルの構築とその応用
完了
小川 芳樹
Understanding individual and crowd dynamics in urban environments are critical for numerous applications, such as urban planning, traffic forecasting, and location-based services. However, researchers have developed travel demand models to accomplish this task with active self-reported data or probe data which are very expensive and often in limited volume. In contrast, emerging data collection methods have enabled researchers to leverage machine learning techniques with a tremendous amount of passively collected mobility data for analyzing and forecasting people’s behaviors. In this study, we plan to develop a reinforcement learning-based approach for modeling and simulation of daily population movement using the Person Trip Survey data. Unlike traditional travel demand modeling approaches, our method focuses on the problem of inferring the spatio-temporal preferences of individuals from the observed trajectories and is based on inverse reinforcement learning (IRL) techniques. We apply the model to the Tokyo area and attempt to replicate a large amount of the population’s daily movement by incorporating with agent-based multi-modal traffic simulation technologies.
変更のために新しい申請を保存します。 This will save a new application on the system for a modification.
申請中の研究者は表示されません。 / Pending researchers are not shown.
龐岩博 / 東京大学 空間情報科学研究センター
申請中のデータセットは表示されません。 / Pending datasets are not shown.
2008年東京都市圏 人の流れデータセット
年次報告の内容はメンバーのみ表示されます。