2015年度

場所: 関西大学第4学舎1号館2階 201A (数学科共通実験実習室)

時間: 14:40 〜 16:10 (I)

講演者: 土田 兼治 氏 (防衛大学校)

講演題目: Large deviations for some additive functionals

時間: 16:20 〜 17:50 (II)

講演者: Liping Li 氏 (中国科学院)

講演題目: Class of smooth functions in Dirichlet spaces

講演要旨: Given a regular Dirichlet form $(\EE,\FF)$ on a fixed domain $E$ of $\mathbb{R}^d$, we first indicate that the basic assumption $C_c^\infty(E)\subset \FF$ is equivalent to the fact that each coordinate function $f^i(x)=x_i$ locally belongs to $\FF$ for $1\leq i\leq d$. In this talk, we will start from these two different viewpoints. On one hand, we shall explore when $C_c^\infty(E)$ is a special standard core of $\FF$ and give some useful characterizations. On the other hand, we shall describe the Fukushima's decompositions of $(\EE,\FF)$ with respect to the coordinates functions, especially discuss when their martingale part is a standard Brownian motion and what we can say about their zero energy part. Finally, when we put these two kinds of discussions together, an interesting class of stochastic differential equations are raised. They have uncountable solutions that do not depend on the initial condition.


場所: 関西大学第4学舎1号館2階 201A (数学科共通実験実習室)

時間: 14:40 〜 16:10 (I)

講演者: 三浦 佑介 氏 (東北大学)

講演題目: The conservativeness of Girsanov transformed symmetric Markov processes

講演要旨: 対称マルコフ過程に対して,対称性が保たれるようなギルサノフ変換を考える.元の過程が連続であるとき, 上の変換により得られる過程が保存的となる条件が既に知られている. 本講演では,元の過程がジャンプを持つ場合でも, 同様の条件の下で変換過程が保存性を持つことを証明する.

時間: 16:20 〜 17:50 (II)

講演者: 塩沢 裕一 氏 (岡山大学)

講演題目:Rate functions for symmetric Markov processes via heat kernel

講演要旨: 本講演は Jian Wang 氏 (Fujian Normal University) との共同研究に基づく。 本講演では,対称マルコフ過程の熱核評価に関する仮定の下で, 与えられた関数が upper/lower rate function となるための積分判定法を得る。 ここで upper/lower rate function とは, 対称マルコフ過程の保存性/(非)再帰性を定量的に表現した関数のことである。


場所: 関西大学第4学舎1号館2階 201A (数学科共通実験実習室)

時間: 15:00 〜 16:30

講演者: 盛田 健彦 氏 (大阪大学)

講演題目: Limit theorems for a class of dynamical systems with quasi-compact Perron-Frobenius operators


場所: 関西大学第4学舎1号館2階 201A (数学科共通実験実習室)

時間: 15:00 〜 16:30

講演者: Sergey Nadtochiy 氏 (University of Michigan)

講演題目: WEAK REFLECTION PRINCIPLE FOR MARKOV PROCESSES

講演要旨: The classical Reflection Principle allows one to express the joint distribution of a Brownian motion and its running maximum through the distribution of the process itself. It relies on the specific symmetry and continuity properties of a Brownian motion and, therefore, cannot be directly applied to an arbitrary Markov process. I will show that, in fact, there exists a weak formulation of this method that allows to obtain similar results for Markov processes that do not posses any strong symmetry. I will describe the general Weak Reflection Principle and will prove its validity for diffusions and Levy processes. I will also demonstrate the applications of this technique in Finance, Computational Methods, and Inverse Problems.