Infinite variety of thermodynamic speed limits with… — Plain-Language Explanation

Imagine you are trying to move a crowd from one side of a room to the other. You want to do this as quickly as possible, but also without wasting too much energy (such as by shouting, pushing, or running in circles). In the world of physics, this is somewhat like shifting a system from one state to another. The "wasted energy" is called entropy production, and the "speed" indicates how fast the system changes.

For a long time, physicists knew a simple rule: you cannot move quickly without wasting some energy. This is the "thermodynamic speed limit." However, until recently, scientists mainly used only one or two specific methods to measure how "busy" or "active" a system is in order to calculate this limit. It was as if one tried to measure a car's speed using only a speedometer while ignoring the engine's RPM or fuel consumption.

This work by Nagayama, Yoshimura, and Ito says: "Wait, there are infinitely many ways to measure this 'busyness'!"

Here is a breakdown of their discovery using simple analogies:

1. The "Activity" of the System

Imagine a busy highway.

The Traffic: The moving cars are the particles in the system.
The Activity: This is a measure of how strongly the cars are moved back and forth.
- In the past, scientists measured activity mainly by simply counting how many cars passed a point (the "Arithmetic Mean").
- Another group measured it by considering the harmony between cars moving forward and backward (the "Logarithmic Mean").

The authors realized that "counting" and "harmony" are just two specific ways to average numbers. In fact, there are infinitely many other ways to average numbers (such as geometric means, contraharmonic means, etc.). They call these different possibilities "General Activities."

2. The Infinite Variety of Speed Limits

The work proves that for each single one of these infinitely many ways to measure "busyness," a new, valid rule for the speed limit can be created.

The Analogy: Imagine a rule stating: "To run one mile in 10 minutes, you must burn 100 calories."
- If you measure "effort" by the number of steps, you get one calorie limit.
- If you measure "effort" by the number of heartbeats, you get a different calorie limit.
- If you measure "effort" by the amount of sweat, you get yet another limit.
- All these limits are true, but they yield different numbers depending on how you define "effort."

The authors show that you can choose any mathematical "mean" to define the system's activity, and you will obtain a valid thermodynamic speed limit. This creates an "infinite variety" of rules.

3. Which Limit is the Best?

You might ask: "If there are infinitely many rules, which one is the strictest? Which one tells me the absolute minimum energy I must waste?"

The answer from the work is surprising: It depends on the situation.

Sometimes the rule based on "counting steps" (Arithmetic Mean) is the strictest.
Sometimes the rule based on "heartbeats" (Logarithmic Mean) is the strictest.
Sometimes a strange, obscure rule (like the "Contraharmonic Mean") is the strictest.

There is no single "best" ruler for all situations. The strictest limit changes depending on how fast the system moves and how far it is from a quiet, balanced state.

4. The Secret of the "Conservative Force"

The work also discovered something beautiful about the perfect path of motion.
If you want to move a system from point A to point B using the absolute minimum amount of wasted energy, there is a specific method to do so. The authors found that this "perfect path" can always be achieved by a conservative force.

The Analogy: Imagine a hiker walking down a mountain. A "conservative force" is like gravity. If you simply let gravity pull you down a smooth path, you waste no energy fighting friction or taking wrong turns.
The work proves that regardless of which "activity" ruler you use to measure the system, the most efficient path is always one that acts like a smooth, gravity-driven slide. You do not need to add extra, chaotic forces to achieve the theoretical minimum energy cost.

5. What About "Excess" Energy?

Sometimes a system is already moving (like a flowing river). The "total" wasted energy includes the energy needed just to keep the river flowing, plus the additional energy needed to change its speed.

The work found that while each of their infinitely many rules works for the total energy waste, only specific rules work for the additional (excess) energy waste.
It is as if one were to say: "Any ruler can measure the total length of a road, but only a certain type of ruler can accurately measure the new pavement you just added."

Summary

In short, this work unites many different thermodynamic speed limits into a single, massive framework.

Infinite Rules: There is not just one speed limit; there are infinitely many valid ones, depending on how you measure the system's "activity."
No Single Winner: No single rule is always the best; the "strictest" limit changes depending on the system's behavior.
The Perfect Path: The most energy-efficient way to move a system is always achievable with a smooth, conservative force, regardless of which rule you use.

The authors did not just find a new rule; they built a toolbox with infinitely many rules, showing us that the laws of thermodynamics are much more flexible and diverse than we previously thought.

Technical Summary: Infinite Variety of Thermodynamic Speed Limits with General Activities

Problem Statement
Thermodynamic speed limits (TSLs) establish fundamental trade-offs between the speed of time evolution, the entropy production rate (EPR), and the kinetic activity of a system. While existing TSLs for Markov jump processes (MJPs) and chemical reaction networks (CRNs) have been derived using specific measures of kinetic intensity—primarily dynamical activity (based on the arithmetic mean of forward and backward flows) and dynamical state mobility (based on the logarithmic mean)—it remains unclear whether these results can be unified or extended to a broader class of "activities." In particular, it is unknown whether general means (such as the geometric mean or other Stolarsky means) can also yield valid TSLs and whether a hierarchy exists among the bounds provided by different means. Furthermore, the conditions under which these bounds are sharp and achievable by conservative forces require a unified theoretical framework.

Methodology
The authors develop a unified framework for deriving TSLs by generalizing the definition of "activity" via homogeneous symmetric means.

General Activity Definition: The paper defines an edge-wise activity $\mu_{m,e}$ for a transition $e$ as the mean $m$ of the forward ( $J^+_e$ ) and backward flows ( $J^-_e$ ): $\mu_{m,e} = m(J^+_e, J^-_e)$ . The total activity is the sum over all edges. This generalizes dynamical activity (arithmetic mean $A$ ) and dynamical state mobility (logarithmic mean $L$ ).
Physical Conditions on Means: To derive TSLs, the authors impose two mathematical conditions on the representing function $f_m(r)$ $f_{m} (r)$ of the mean $m$ $m$ :
- Condition (24): Ensures that the function $\Psi_m(u) = (e^u - 1)/f_m(e^u)$ is monotonically increasing, enabling a bijective mapping between current and force (a nonlinear extension of the Onsager reciprocal relation).
- Condition (25): Ensures that the convexity of the function $w\Psi^{-1}_m(w)$ is given, which is essential for establishing the monotonicity of the EPR under coarse-graining.
  The authors prove that broad families of means, including the Stolarsky mean and the contraharmonic mean, satisfy these conditions.
Derivation of TSLs: Using the 1-Wasserstein distance ( $W_1$ ) to measure the speed of time evolution ( $v_1$ ), the authors derive a general trade-off relation. By applying Jensen's inequality to the convex function $w\Psi^{-1}_m(w)$ , they establish a lower bound for the time-averaged EPR ( $\langle \sigma \rangle_\tau$ ) in terms of the time-averaged speed and activity:
$\langle \sigma \rangle_\tau \ge \langle v_1 \rangle_\tau \Psi^{-1}_m \left( \frac{\langle v_1 \rangle_\tau}{\langle \mu_m \rangle_\tau} \right)$
This formulation encompasses previously known TSLs as special cases.
Analysis of Sharpness and Hierarchy: The authors investigate whether a hierarchy exists among TSLs derived from different means (i.e., whether $m_1 \le m_2$ implies a sharper bound). They analyze this both analytically (in low-speed regimes and near equilibrium) and numerically for various system states.
Minimal Dissipation and Achievability: The paper formulates a minimization problem for entropy production subject to a constraint on time-averaged activity. It is shown that the lower bound in the TSL is achievable and corresponds to a formula for minimal dissipation. The optimal protocol involves a time-independent current and a conservative force.
Comparison with Excess EPR: The authors compare their general TSLs with bounds for the excess entropy production rate (specifically the Onsager geometric and information-geometric excess EPRs). They contrast the behavior of the bounds for the total EPR with that of the bounds for the excess EPR.

Main Contributions and Results

Infinite Variety of TSLs: The paper derives an infinite family of TSLs by employing general homogeneous symmetric means (e.g., Stolarsky means) as activities. This unifies existing TSLs based on arithmetic and logarithmic means.
Unified Framework for Minimal Dissipation: The authors prove that for any valid mean $m$ , the minimal dissipation required to evolve a system between two states within time $\tau$ is given by the lower TSL bound. Crucially, this minimum is always achievable by a conservative force and a time-independent current, regardless of the specific mean chosen as the activity. This generalizes previous results restricted to specific means.
Hierarchy of Bounds:
- General Regime: Numerical results show that no universal hierarchy exists among TSLs; the sharpness of the bound depends on the specific system state and time. A smaller mean generally does not yield a sharper bound in non-stationary states.
- Specific Regimes: Analytical results confirm that for certain pairs of means, a hierarchy exists (e.g., the TSL based on the minimum mean is always sharper than or equivalent to that based on the maximum mean). Additionally, in low-speed regimes (near stationary states), smaller means yield sharper bounds.
Excess EPR Constraints: The study reveals a critical difference between bounds for the total and the excess EPR. While every valid mean provides a lower bound for the total EPR, only specific means (arithmetic and logarithmic) provide valid lower bounds for the excess EPR. TSLs based on other means (e.g., geometric or harmonic) may fail to bound the excess EPR, suggesting that the choice of mean is crucial when analyzing non-stationary contributions.
Comparison with the 2-Wasserstein Distance: The authors compare their 1-Wasserstein-distance-based TSLs with those based on the 2-Wasserstein distance (which bounds the Onsager excess EPR). They note that contrary to behavior in Langevin systems, where the 2-Wasserstein bound is strictly sharper, certain 1-Wasserstein-distance-based TSLs (using specific means) can be sharper than the 2-Wasserstein bound in discrete systems.

Significance and Claims
The paper claims to provide a "unified framework" that explains the origin of various existing TSLs and extends them to an infinite variety of new bounds. The significance lies in:

Unification: It is shown that the difference between dynamical activity and dynamical state mobility is merely a choice of averaging mean, and that a broad class of means can serve as valid activities for deriving TSLs.
Achievability: It is established that the lower bounds are not merely theoretical limits but are achievable by conservative forces, providing a general class of conditions for minimal dissipation.
Nuance in Excess EPR: It is highlighted that the validity of TSLs for excess entropy production is more restrictive than for total entropy production, suggesting that the definition of excess EPR is inherently linked to the choice of mean (geometry) used in its construction.
No Universally Best Bound: The work challenges the notion of a single "best" TSL, demonstrating that the sharpest bound is context-dependent and varies with the system's distance from equilibrium and the speed of evolution.

The authors modestly note that although they have characterized the conditions for these bounds, exploring the correspondence between the nature of the dynamics and the specific mean that yields the sharpest bound remains an open challenge. They also identify extending these results to open quantum systems and clarifying the relationship with classical information-geometric speed limits as future directions.

Infinite variety of thermodynamic speed limits with general activities