Title: Thermodynamic Variational Structure and Optimal Control of Chemical Reaction
Speaker: Prof. Tetsuya J. Kobayashi (Institute of Industrial Science, the University of Tokyo)
Host: Prof. Cheng-Hung Chang (NYCU)
Time: 14:00, September 24 (THU.), 2026
Place: Rm. 307, 3F, Cosmology Hall, NTU

Abstract:
The variational representation of dynamical systems is a powerful framework that unveils a system's hidden properties, geometric aspects, and underlying universality. 
In classical mechanics, expressing Newton's equations through the variational lens of Lagrange's equations have revealed their intrinsic geometric structure and physical universality, which extends far beyond mechanics itself. 

Driven by recent advances in non-equilibrium thermodynamics, geometric analysis, and large deviation theory, variational representations are now being uncovered in non-equilibrium chemical reaction networks as well. Crucially, owing to its deep link with large deviations, this representation implicitly captures not just deterministic dynamics, but also the system's stochastic behavior. 

In this talk, I will overview the variational representations and information-geometric structures of non-equilibrium chemical reaction networks[1], demonstrating how this framework offers a unified path to deriving thermodynamic uncertainty relations, speed limits, and related bounds[2]. Finally, I will discuss how this variational structure can be leveraged to tackle control problems in non-equilibrium chemical networks[3].

[1] T.J. Kobayashi et al, Phys. Rev. Res., 4, 033208 (2022); 4, 033066 (2022); T.J. Kobayashi, et al,"Information geometry of dynamics on graphs and hypergraphs", Info. Geom. 7, 97-166 (2024)
[2] D. Loutchko et al, "Information geometry of perturbed gradient flow systems on hypergraphs: A perspective towards nonequilibrium physics", arXiv:2510.27268 (2025) to appear in Info. Geom.
[3] S.A. Horiguchi & T.J. Kobayashi, "Optimal Control of Stochastic Reaction Networks with Entropic Control Cost and Emergence of Mode-Switching Strategies", PRX Life 3, 033027 (2025)