発表者名 |
東條 開斗 |
指導教員名 |
沙川 貴大 教授 |
発表題目(英語) |
Optimization of finite-time work extraction and its application to information-to-work conversion |
要旨(英語) |
Maximizing extractable work from a given system is crucial in designing thermodynamic machines, such as heat engines and biomolecular systems. While the theoretical upper bound of extractable work is quantified by the second law of thermodynamics, achieving this bound requires a quasi-static process, rendering it unattainable in finite-time operations. Thus, maximizing extractable work in finite time is a fundamental challenge. Recent studies have refined the second law by utilizing optimal transport theory, quantifying the maximal amount of extractable work within finite time for fixed initial and final states. However, the choice of final state can be further optimized to maximize extractable work, making it unnatural to fix the final state when aiming to maximize extractable work. In particular, for two coupled systems involving information processing, extractable work is enhanced beyond conventional second law utilizing correlations between the systems. Consequently, the methods for maximizing work extraction utilizing correlations from a given initial state through finite-time operations have not yet been well established.
In this study, we derive a method to maximize work extraction in finite-time processes by employing optimal transport theory and variational methods. Our general results are applicable to non-Gaussian processes,
while for Gaussian processes, we exploit their specific properties to derive an analytical expression of the extractable work. We demonstrate our framework by considering the setup of Maxwell’s demon, which performs feedback based on measured information of the system. Furthermore, we examine its information- thermodynamic efficiency and discuss its relation to the thermodynamic operational characterization of correlation in finite-time processes. |
発表言語 |
日本語 |