Markov chains are a fundamental subject of study in mathematical probability and have found wide application in nearly every branch of science. Of particular interest are finite-state Markov chains; the representation of finite-state Markov chains by a transition matrix facilitates detailed analysis by linear algebraic methods. Previous methods of analyzing finite-state Markov chains have emphasized state events. In this thesis we develop the concept of a transition event and define two types of transition events: cumulative events and time-average events. Transition events generalize state events and provide a more flexible framework for analysis. We derive computable, closed-form expressions for the expectation of these two events, characterize the conditioning of transition events, provide an algorithm for computing the expectation of these events, and analyze the complexity and stability of the algorithm. As an application, we derive a construction of composite Markov chains, which we use to study competitive dynamics.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21397号 / 情博第683号 / 新制||情||118(附属図書館) / 京都大学大学院情報学研究科社会情報学専攻 / (主査)教授 吉川 正俊, 教授 黒田 知宏, 教授 守屋 和幸 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM