Let , and first suppose . On any transport plan , which has total mass one, the Holder inequality gives
Also, gives
Taking infima, or using arbitrarily close competitors if an infimum is not attained, yields
For the two distances coincide, and if there is only one probability measure, so the conclusion is immediate. For and , the displayed bounds imply both directions of
For , exchange the exponents. Thus all finite-order p-Wasserstein distances induce the same convergence on a bounded subset of . Closedness of is not required for these estimates.
For , define the probability measures with finite th absolute moment by
where is any fixed reference point. The condition is independent of the choice of . The p-Wasserstein distance is
The infimum is over all transport plans, rather than only transport maps. The product measure gives a finite upper bound using
For this is the Wasserstein distance with the metric of Euclidean space; for it is the second Wasserstein distance. “Bounded moment” here means a finite integral, and does not require to be bounded.
Define ; the printed quotient is undefined at the origin, but this continuous extension changes no transport cost because . In polar coordinates, the two probability density functions give
Both angular distributions are uniform, with independence of angle and radius. The proposed transport map preserves the angle and sends to . Therefore
and preservation of the angle proves . As a local check using the Jacobian determinant, its radial and tangential derivatives for are and , respectively, so and .
Now use the convex function
It is convex because the Euclidean norm is convex and is increasing and convex on . It is differentiable, including at zero, and
Its graph transport plan lies in the graph of , so the Knott–Smith optimality criterion proves quadratic optimality. Since that plan is induced by a map, part 1(b) proves optimality for the Monge optimal transport problem as well. The minimum provides a useful independent check:
Let and let denote its convex conjugate. Its Fenchel–Young gap
is nonnegative by the Fenchel–Young inequality, and the hypothesis says .
First justify the integrability needed for the certificate. Finite second moments and the Cauchy-Schwarz inequality imply, for every transport plan ,
Since and , the identity proves by the marginal distribution property. In particular, no subtraction of infinite integrals is being used.
Set
These are integrable Kantorovich potentials. The Fenchel–Young inequality gives . More precisely, for every transport plan ,
Taking the infimum in the first inequality therefore proves
The factor in the quadratic cost is essential for this exact gap identity.
A standard sufficient form of Brenier theorem assumes have finite second moments and , that is, absolute continuity of measures with respect to Lebesgue measure. For the cost , there exists a unique optimal transport plan, and it is induced by a transport map:
Here is a proper convex function, which may be chosen sequentially lower semicontinuous, and its gradient exists -almost everywhere. The map is unique -almost everywhere and is the unique minimizer of the Monge optimal transport problem; its cost equals the Kantorovich optimal transport problem minimum. Equivalently, it is the unique gradient of a convex function transporting to .
The uniqueness claim concerns the map and the transport plan, not a globally unique potential. The potential may be shifted by a constant, and additional nonuniqueness away from the source can occur. No density assumption is required on . A primary reference is Brenier's Polar factorization and monotone rearrangement of vector-valued functions.
For probability measures on with finite second moments, the Knott–Smith optimality criterion states that a transport plan minimizes the quadratic cost if and only if there is a proper convex function that is sequentially lower semicontinuous and satisfies
The subdifferential is characterized by
with . Thus the transport plan is concentrated on the graph of the subdifferential. No absolute continuity of measures assumption on is needed. Multiplying the cost by leaves the criterion unchanged. The original quadratic optimal mapping result is Knott and Smith, On the optimal mapping of distributions.
Put and use the two Kantorovich potentials
For every ,
Thus the pair is dual feasible, with equality precisely on . Finite second moments make both potentials integrable; compactness of is more than sufficient. For any admissible transport map , integrating the inequality and using its pushforward measure gives
The admissible map attains equality. Consequently
The same certificate proves optimality of its graph transport plan among all transport plans. Notice that need not be a convex function when : it is a cost dual potential. The associated convex gradient potential is instead .
The Kantorovich–Rubinstein theorem concerns a Polish space and probability measures with finite first moments, meaning for one, hence every, . It states
Here means for every . The Wasserstein distance with ground cost therefore equals a supremum over functions with Lipschitz constant at most one. One may normalize , since adding a constant does not change the difference of integrals. This normalization bounds by , so finite first moments ensure integrability. The supremum is unchanged if an absolute value is placed around the difference, because is admissible whenever is.
Fix any admissible pair of Kantorovich potentials and any transport plan . Its marginal distributions give
The right side is well defined because and are Lebesgue integrable with respect to . Integrating their pointwise feasibility inequality gives
Since this holds for every feasible pair and every transport plan,
This proves the required inequality directly from the transport constraints, without any convex optimization duality theorem. Whenever the extrema are attained, the supremum and infimum can respectively be written as a maximum and minimum. Even the Kantorovich duality theorem is unnecessary for this direction.
The dual of the Kantorovich optimal transport problem is
where the Kantorovich potentials are measurable representatives satisfying
One standard form of the Kantorovich duality theorem assumes that are Polish spaces, are probability measures defined as Borel measures, and is sequentially lower semicontinuous. Then
The primal infimum is attained; its value may be . Nonnegativity can be replaced by a constant lower bound by shifting the cost. This duality for lower semicontinuous costs is also discussed in Beiglboeck, Leonard and Schachermayer's duality paper.
Equality of values does not by itself assert a dual maximum. A sufficient stronger setting for attainment on both sides is compact metric spaces and a finite continuous cost ; then continuous Kantorovich potentials attain the dual supremum. The general statement above correctly uses a supremum.
As in part (c), every admissible transport map has nonnegative displacement with . The square root is a strictly concave function, so the reversed Jensen inequality gives
The map is admissible and attains one. Therefore
Thus “worst” means largest cost among admissible transport maps. Equality in the Jensen inequality for a strictly concave function forces to be constant almost everywhere, so this maximizer is unique up to a -null set. For comparison, the admissible reflection has smaller cost
The upper bound also holds for every transport plan.
For every admissible transport map , the pushforward measure condition gives for -almost every . Hence , and
Apply the Jensen inequality to the convex function and the probability measure :
Translation by one sends the source uniform distribution to the target uniform distribution, so is admissible. Its displacement is constantly one, giving
No monotonicity assumption on is needed, because all displacements have the same sign. The same argument with a transport plan also gives the identical minimum for the Kantorovich optimal transport problem.
Write , the absolute Jacobian determinant. Under the usual meaning of a map on domains, the injective area formula gives, for any nonnegative measurable ,
This version of the change of variables formula does not require to be nonsingular everywhere. A map is a locally Lipschitz function, and the formula applies by exhaustion if the domains are unbounded. For non-open , one needs the corresponding extension or regularity interpretation of the stated assumption.
First suppose almost everywhere with respect to Lebesgue measure. For each bounded nonnegative measurable ,
Taking indicator functions proves .
Conversely, the area formula shows that the measure with probability density function has total mass one and satisfies . The assumed equality then implies , since the measurable bijection has a measurable inverse on these domains. Equivalently, apply both pushforward measure identities to for each Borel set . Uniqueness of the Radon-Nikodym derivative yields
Interpret the invertibility assumption on as continuity and strict increase on the relevant range, so that is an atomless measure. Suppose a non-decreasing transport map with exists. Then is also an atomless measure: if , the pushforward measure would give . Thus is continuous.
At any point where the non-decreasing representative is defined, the definition of a monotone function gives
Using the pushforward measure identity and continuity of the two cumulative distribution functions, we obtain
Consequently
The monotone rearrangement in part (c) is optimal for the convex difference cost, so has the same cost and solves the Monge optimal transport problem. The meaningful uniqueness is up to a -null set; arbitrary values away from the source do not affect transport or cost. Existence of the non-decreasing transport map supplies the source condition missing in part (c).
The intended monotone rearrangement is
Here is the quantile function of , agreeing with the ordinary inverse when is continuous and strictly increasing. With an atomless measure , its cumulative distribution function is continuous, and the probability integral transform makes uniform on for . Thus . The one-dimensional monotone rearrangement theorem says this transport map minimizes the cost for convex continuous , whenever the cost integrals are well defined. Values at exceptional endpoints may be chosen arbitrarily.
The printed assumptions omit an essential source condition. Invertibility of alone does not ensure an admissible transport map. For example, and a standard normal distribution satisfy the stated condition on , but is always a Dirac measure. There is no solution to the Monge optimal transport problem in this example. The boxed answer therefore requires the additional assumption that is an atomless measure, or an equivalent condition making the displayed map admissible. For arbitrary sources the always admissible monotone transport plan is , which need not be induced by a map.
Given any admissible transport map , form its graph transport plan
For Borel sets and , the definition of a pushforward measure gives
Hence . Integration against a pushforward measure also gives
The Kantorovich optimal transport problem therefore has at least all the competitors of the Monge optimal transport problem, with exactly the same costs. Consequently
If there is no admissible transport map, the left side is by the convention , so the conclusion still holds. No existence of an optimizer is needed.
For a cost that is a Borel measurable function, a transport map is a measurable whose pushforward measure satisfies
The Monge optimal transport problem moves every source point to one destination:
The Kantorovich optimal transport problem permits mass to split. Its admissible transport plans are the probability measures on with prescribed marginal distributions:
Thus a transport plan is a coupling of probability distributions. The set is never empty: it contains the product measure . Signed costs can also be used when their integrals are well defined, for example with an integrable lower bound of the form .
On the Polish space , take the Dirac measures
Every measurable map satisfies , which cannot equal . A transport map cannot split an atom of a measure, whereas the transport plan can.
Interpret as the rest-mass supply rate through a coherent disc; the hole's gravitating mass grows more slowly because radiation carries away binding energy. Initially the hole is a Schwarzschild black hole, whose innermost stable circular orbit is at . The relativistic orbital constants there are
They follow by extremizing the timelike geodesic effective potential for a circular orbit and imposing marginal stability. Assume negligible torque inside the inner edge, so the gas carries this specific angular momentum into the hole. Its initial spin-up torque is
for and . In SI units this is . A purely Newtonian estimate gives a similar order of magnitude, but the relativistic value is appropriate at the innermost stable circular orbit.
The maximal Kerr black hole angular momentum at the initial mass is . Holding both that mass and the initial torque fixed would give
Initially , so the corresponding frozen-coefficient estimate of gravitating mass gain is about . These are useful initial spin-growth scales, not a self-consistent time and mass increase for reaching extremality: the target angular momentum itself grows as , and the inner orbit changes as the spin increases.
A consistent endpoint estimate is obtained from black-hole spin-up by coherent accretion. Let and , where is the current mass. On the prograde Kerr black hole inner-orbit branch, , the orbital constants may be written
Ignoring photon capture, an accreted rest mass changes the hole by and . Differentiating the spin definition gives
Substitution of the orbital constants gives , hence
At the formal extremal limit , the result is
For constant rest-mass supply, use to integrate the elapsed time:
Thus the idealized coherent-disc endpoint is
If the quoted were instead the actual gravitating mass-growth rate, the corresponding endpoint time would be . The two rate conventions cannot be interchanged.
The exact extremal endpoint is an idealization. Capture of disc photons exerts a counteracting torque and gives the Thorne spin limit, , for the standard radiatively efficient thin-disc model. Therefore physical disc accretion approaches a large spin below unity rather than producing an exactly extremal hole. Thorne's original spin-evolution calculation describes that correction. The independently integrated expressions above explain why the frozen-mass estimate is too short for the ideal endpoint.
The qualitative conclusion is robust: a hole that gains most of its mass through a persistent, coherently corotating Shakura--Sunyaev thin disk is expected to have substantial black-hole spin. Very low spins require a different angular-momentum history, such as short episodes with changing orientations, counterrotating accretion, or mergers. Coherent feeding, not merely the existence of a disc during each episode, is the decisive assumption.
The Soltan argument offers a population-level observational test. From the bolometric luminosity function of active galactic nuclei, infer the total energy radiated per comoving volume over cosmic history,
Compare it with the increase in supermassive black hole mass density attributed to radiative accretion. For an effective radiative efficiency of black-hole accretion ,
In a zero-torque relativistic thin disc, rises from about at zero spin to in the photon-free prograde extremal limit. The measured effective efficiency therefore constrains population spin models under the thin-disc assumption. Soltan's original population argument explains how integrated quasar light constrains accumulated black-hole mass.
This is an accreted-rest-mass-weighted efficiency constraint, not a direct measurement of the arithmetic mean spin of every black hole. Obscured emission, bolometric corrections, seed masses, nonradiative growth, and disc orientation introduce uncertainty, and the efficiency-spin relation is nonlinear. These qualifications are essential when interpreting an inferred average spin for the whole population.

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact