主催: The Japanese Society for Artificial Intelligence
会議名: 2019年度人工知能学会全国大会（第33回）
回次: 33
開催地: 新潟県新潟市 朱鷺メッセ
開催日: 2019/06/04 - 2019/06/07
The main objective of this paper is to analyze some variants of the classical top-trading-cycles (TTC) algorithm for slightly modified models of the housing market. Extensions of TTC for such modified models are not necessarily strategy-proof, as pointed out by Fujita et al.\ (2015), and thus some alternative analysis of agents' selfish behavior is needed. In this paper, the incentive ratio, originally proposed by Chen et al.\ (2011), of the variants of TTC algorithm is analyzed in both (i) the multi-item exchange and (ii) an exchange model with a specific form of externalities.
