报告题目：Toward an Axiomatic Foundation of Nonequilibrium Statistical Mechanics

报告人：吴建中教授，美国加州大学河滨分校

报告时间：2026年10月27日（周二）14：30

报告地点：电化学楼C512会议室

报告人简介：吴建中教授，现任美国加州大学河滨分校化学与环境工程杰出教授。本科、硕士均毕业于清华大学，博士毕业于美国加州大学伯克利分校。主要研究方向为经典密度泛函理论、统计热力学、多尺度模拟及其在交叉学科问题中的应用。他是美国物理学会（APS）、美国化学工程师学会（AIChE）以及美国医学与生物工程91社区 （AIMBE）会士，在非平衡统计力学、分子热力学领域有重要国际影响力。

报告简介：How do macroscopic irreversibility and transport emerge from microscopic dynamics? Motivated by Hilbert’s sixth problem, this seminar explores a unified foundation for nonequilibrium statistical mechanics built around six principles: probability, Gibbs–Shannon entropy, maximum-entropy inference, microscopic reversibility, symmetry, and dynamical correlations. The first part develops a probabilistic description of physical states and examines the equivalence between microstate and trajectory representations of Gibbs–Shannon entropy. Physically specified forward and reverse path measures connect this description to entropy production, fluctuation relations, and the second law. The discussion distinguishes microscopic entropy from trajectory-dependent stochastic entropy and explains how microscopic reversibility is compatible with irreversible behavior in reduced descriptions. The second part addresses the construction of constitutive equations. Starting from Hamiltonian and Langevin dynamics, hierarchies of reduced distributions yield exact balance equations for particle number, momentum, and energy. A pair-level maximum-entropy closure provides a constructive route toward fluid dynamics. Extensions that retain relative pair currents and energy transport reveal the correlations required beyond a local shifted-Gaussian description. The proposed foundation connects Boltzmann’s kinetic theory, Onsager’s irreversible thermodynamics, and stochastic thermodynamics, highlighting the role of dynamical correlations in extending constitutive descriptions beyond linear response.