# Why is Entropy maximised when the probability distribution is uniform?

I know that entropy is the measure of randomness of a process/variable and it can be defined as follows. for a random variable $X \in$ set $A$ :- $H(X)= \sum_{x_i \in A} -p(x_i) \log (p(x_i))$. In the book on Entropy and Information Theory by MacKay, he provides this statement in Ch2

Entropy is maximized if p is uniform.

Intuitively, I am able to understand it, like if all datapoints in set $A$ are picked with equal probability $1/m$ ($m$ being cardinality of set $A$), then the randomness or the entropy increases. But if we know that some points in set $A$ are going to occur with more probability than others (say in the case of normal distribution, where the maximum concentration of data points is around the mean and small standard deviation area around it, then the randomness or entropy should decrease.

But is there any mathematical proof for this ? Like the equation for $H(X)$ I differentiate it with respect to $p(x)$ and set it to 0 or something like that.

On a side note, is there any connnection between the entropy that occurs information theory and the entropy calculations in chemistry (thermodynamics) ?

• This question is answered (in passing) at stats.stackexchange.com/a/49174/919.
– whuber
Jul 31, 2013 at 14:52
• I am getting quite confused with another statement given in Christopher Bishops book which states that "for a single real variable, the distribution that maximizes the entropy is the Gaussian." It also states that "multivariate distribution with max- imum entropy, for a given covariance, is a Gaussian". How is this statement valid? Isnt the entropy of the uniform distribution the maximum always? Sep 5, 2014 at 7:16
• Maximization is always performed subject to constraints on the possible solution. When the constraints are that all probability must vanish beyond predefined limits, the maximum entropy solution is uniform. When instead the constraints are that the expectation and variance must equal predefined values, the ME solution is Gaussian. The statements you quote must have been made within particular contexts where these constraints were stated or at least implicitly understood.
– whuber
Sep 5, 2014 at 15:57
• I probably also should mention that the word "entropy" means something different in the Gaussian setting than it does in the original question here, for then we are discussing entropy of continuous distributions. This "differential entropy" is a different animal than the entropy of discrete distributions. The chief difference is that the differential entropy is not invariant under a change of variables.
– whuber
Sep 5, 2014 at 17:28
• So which means that maximisation always is with respect to constraints ? What if there are no constraints ? I mean, cant there be a question like this ? Which probability distribution has maximum entropy ? Feb 17, 2016 at 10:44

Heuristically, the probability density function on $\{x_1, x_2,..,.x_n\}$ with maximum entropy turns out to be the one that corresponds to the least amount of knowledge of $\{x_1, x_2,..,.x_n\}$, in other words the Uniform distribution.

Now, for a more formal proof consider the following:

A probability density function on $\{x_1, x_2,..,.x_n\}$ is a set of nonnegative real numbers $p_1,...,p_n$ that add up to 1. Entropy is a continuous function of the $n$-tuples $(p_1,...,p_n)$, and these points lie in a compact subset of $\mathbb{R}^n$, so there is an $n$-tuple where entropy is maximized. We want to show this occurs at $(1/n,...,1/n)$ and nowhere else.

Suppose the $p_j$ are not all equal, say $p_1 < p_2$. (Clearly $n\neq 1$.) We will find a new probability density with higher entropy. It then follows, since entropy is maximized at some $n$-tuple, that entropy is uniquely maximized at the $n$-tuple with $p_i = 1/n$ for all $i$.

Since $p_1 < p_2$, for small positive $\varepsilon$ we have $p_1 + \varepsilon < p_2 -\varepsilon$. The entropy of $\{p_1 + \varepsilon, p_2 -\varepsilon,p_3,...,p_n\}$ minus the entropy of $\{p_1,p_2,p_3,...,p_n\}$ equals

$$-p_1\log\left(\frac{p_1+\varepsilon}{p_1}\right)-\varepsilon\log(p_1+\varepsilon)-p_2\log\left(\frac{p_2-\varepsilon}{p_2}\right)+\varepsilon\log(p_2-\varepsilon)$$ To complete the proof, we want to show this is positive for small enough $\varepsilon$. Rewrite the above equation as $$-p_1\log\left(1+\frac{\varepsilon}{p_1}\right)-\varepsilon\left(\log p_1+\log\left(1+\frac{\varepsilon}{p_1}\right)\right)-p_2\log\left(1-\frac{\varepsilon}{p_2}\right)+\varepsilon\left(\log p_2+\log\left(1-\frac{\varepsilon}{p_2}\right)\right)$$

Recalling that $\log(1 + x) = x + O(x^2)$ for small $x$, the above equation is $$-\varepsilon-\varepsilon\log p_1 + \varepsilon + \varepsilon \log p_2 + O(\varepsilon^2) = \varepsilon\log(p_2/p_1) + O(\varepsilon^2)$$ which is positive when $\varepsilon$ is small enough since $p_1 < p_2$.

A less rigorous proof is the following:

Consider first the following Lemma:

Let $p(x)$ and $q(x)$ be continuous probability density functions on an interval $I$ in the real numbers, with $p\geq 0$ and $q > 0$ on $I$. We have $$-\int_I p\log p dx\leq -\int_I p\log q dx$$ if both integrals exist. Moreover, there is equality if and only if $p(x) = q(x)$ for all $x$.

Now, let $p$ be any probability density function on $\{x_1,...,x_n\}$, with $p_i = p(x_i)$. Letting $q_i = 1/n$ for all $i$, $$-\sum_{i=1}^n p_i\log q_i = \sum_{i=1}^n p_i \log n=\log n$$ which is the entropy of $q$. Therefore our Lemma says $h(p)\leq h(q)$, with equality if and only if $p$ is uniform.

Also, wikipedia has a brief discussion on this as well: wiki

• I admire the effort to present an elementary (Calculus-free) proof. A rigorous one-line demonstration is available via the weighted AM-GM inequality by noting that $\exp(H)$ = $\prod\left(\frac{1}{p_i}\right)^{p_i}\le\sum p_i\frac{1}{p_i}=n$ with equality holding iff all the $1/p_i$ are equal, QED.
– whuber
Jul 31, 2013 at 15:05
• I don't understand how $\sum{\log{n}}$ can be equal to $\log{n}$. Aug 3, 2016 at 14:24
• @user1603472 do you mean $\sum\limits_{i=1}^n p_i \log n = \log n$? Its because $\sum\limits_{i=1}^n p_i \log n = \log n \sum\limits_{i=1}^n p_i = \log n \times 1$ Sep 29, 2016 at 11:18
• @Roland I pulled the $\log n$ outside of the sum since it does not depend on $i$. Then the sum is equal to $1$ because $p_1,\ldots,p_n$ are the densities of a probability mass function. Sep 18, 2017 at 17:57
• Same explanation with more details can be found here: math.uconn.edu/~kconrad/blurbs/analysis/entropypost.pdf Sep 23, 2017 at 10:19

Entropy in physics and information theory are not unrelated. They're more different than the name suggests, yet there's clearly a link between. The purpose of entropy metric is to measure the amount of information. See my answer with graphs here to show how entropy changes from uniform distribution to a humped one.

The reason why entropy is maximized for a uniform distribution is because it was designed so! Yes, we're constructing a measure for the lack of information so we want to assign its highest value to the least informative distribution.

Example. I asked you "Dude, where's my car?" Your answer is "it's somewhere in USA between Atlantic and Pacific Oceans." This is an example of the uniform distribution. My car could be anywhere in USA. I didn't get much information from this answer.

However, if you told me "I saw your car one hour ago on Route 66 heading from Washington, DC" - this is not a uniform distribution anymore. The car's more likely to be in 60 miles distance from DC, than anywhere near Los Angeles. There's clearly more information here.

Hence, our measure must have high entropy for the first answer and lower one for the second. The uniform must be least informative distribution, it's basically "I've no idea" answer.

The mathematical argument is based on Jensen inequality for concave functions. That is, if $f(x)$ is a concave function on $[a,b]$ and $y_1, \ldots y_n$ are points in $[a,b]$, then: $n \cdot f(\frac{y_1 + \ldots y_n}{n}) \geq f(y_1) + \ldots + f(y_n)$

Apply this for the concave function $f(x) = -x \log(x)$ and Jensen inequality for $y_i = p(x_i)$ and you have the proof. Note that $p(x_i)$ define a discrete probability distribution, so their sum is 1. What you get is $log(n) \geq \sum_{i=1}^n - p(x_i) log(p(x_i))$, with equality for the uniform distribution.

• I actually find the Jensen's inequality proof to be a much deeper proof conceptually than the AM-GM one. Oct 8, 2018 at 17:17

On a side note, is there any connnection between the entropy that occurs information theory and the entropy calculations in chemistry (thermodynamics) ?

Yes, there is! You can see the work of Jaynes and many others following his work (such as here and here, for instance).

But the main idea is that statistical mechanics (and other fields in science, also) can be viewed as the inference we do about the world.

As a further reading I'd recommend Ariel Caticha's book on this topic.

Main idea: take partial derivative of each $$p_i$$, set them all to zero, solve the system of linear equations.

Take a finite number of $$p_i$$ where $$i=1,...,n$$ for an example. Denote $$q = 1-\sum_{i=0}^{n-1} p_i$$.

\begin{align} H &= -\sum_{i=0}^{n-1} p_i \log p_i - (1-q)\log q\\ H*\ln 2 &= -\sum_{i=0}^{n-1} p_i \ln p_i - (1-q)\ln q \end{align} \begin{align} \frac{\partial H}{\partial p_i} &= \ln \frac{q}{p_i} = 0 \end{align} Then $$q = p_i$$ for every $$i$$, i.e., $$p_1=p_2=...=p_n$$.

• I am glad you pointed out this is the "main idea," because it's only a part of the analysis. The other part--which might not be intuitive and actually is a little trickier--is to verify this is a global minimum by studying the behavior of the entropy as one or more of the $p_i$ shrinks to zero.
– whuber
Mar 4, 2019 at 22:20

There are already several good answers. Another argument uses the fact that H is a symmetric, strictly concave function. More precisely, consider the unit simplex $$\Delta_n=\{(p_1,\dots,p_n): p_i\ge 0,\sum_i p_i=1\}$$.Then $$H$$ may be considered a function $$H: \Delta_n\to \mathbb{R}$$, and it is easy to show that it is strictly convex.

To explain symmetry, we first introduce some notation. Given a permutation $$\sigma: \{1,\dots, n\}\to\{1,\dots,n\}$$, and a point $$p\in \Delta_n$$, define $$\sigma p=(p_{\sigma(1)},\dots, p_{\sigma(n)})$$. It is clear that $$H(\sigma p)=H(p)$$ for any $$\sigma$$ and p, and this is what it means to say that H is a symmetric function.

Now, we can show that the uniform distribution maximizes $$H$$. Since $$H$$ is strictly convex, it has a unique maximizer, call it $$p_{max}$$. On the other hand $$H(\sigma p_{max})=H(p_{max})$$ for any $$\sigma$$, so $$\sigma p_{max}$$ is also a maximizer. Since $$p_{max}$$ is the only maximizer, we conclude $$p_{max}=\sigma p_{max}$$ for each $$\sigma$$, and the only point in $$\Delta_n$$ with this property is $$p_{max}=(1/n,\dots, 1/n)$$.

Calculus of Variations

To handle varying functions, we will make use of the Calculus of Variations. The variation $$\delta f(x)$$ refers to a rate of change of $$f(x)$$ with respect to "time". That is, $$\delta$$ works like a partial derivative with respect to "time". For example, \begin{align} \delta(\log(f(x))f(x)) &=\left(\frac1{f(x)}f(x)+\log(f(x))\right)\delta f(x)\\ &=(1+\log(f(x)))\,\delta f(x) \end{align} simply says that the rate of change of $$\log(f(x))f(x)$$ is $$(1+\log(f(x)))$$ times the rate of change of $$f(x)$$.

Maximize Entropy

Consider the family of continuous probability distributions $$f$$ on $$[a,b]$$; that is, positive $$f$$ where $$\int_a^bf(x)\,\mathrm{d}x=1\tag1$$ Define the entropy of $$f$$ to be $$-\int_a^b\log(f(x))f(x)\,\mathrm{d}x\tag2$$ If we wish to maximize $$(2)$$ for all distributions satisfying $$(1)$$, we need to find all $$f$$ so that $$(2)$$ is stationary; that is, $$\delta$$ of the integral in $$(2)$$ vanishes: $$-\int_a^b(1+\log(f(x))\,\delta f(x)\,\mathrm{d}x=0\tag3$$ for all variations, $$\delta f$$, where $$(1)$$ is stationary; that is, $$\delta$$ of the integral in $$(1)$$ vanishes: $$\int_a^b\color{#C00}{1}\,\delta f(x)\,\mathrm{d}x=0\tag4$$ $$(3)$$, $$(4)$$, and orthogonality (equations $$(3)$$ and $$(4)$$ say that $$\color{#C00}{1}$$ and $$1+\log(f(x))$$ are orthogonal to all the same variations; see the link for details) require that there be a $$c_0$$ so that $$1+\log(f(x))=c_0\cdot\color{#C00}{1}\tag5$$ That is, the desired distribution is constant; that is, $$\bbox[5px,border:2px solid #C0A000]{f(x)=\frac1{b-a}}\tag6$$

Vanishing Densities

Note that in $$(3)$$, $$-(1+\log(f(x)))\to\infty$$ as $$f(x)\to0$$. This doesn't cause a problem in $$(2)$$ since $$-\log(f(x))f(x)$$ is bounded by $$\frac1e$$. There are a few simple ways to overcome the problem in $$(3)$$.

1. We can use $$\delta f(x)=f(x)\,\delta\log(f(x))$$ and use variations of $$\log(x)$$. Then $$(5)$$ becomes $$(1+\log(f(x)))f(x)=c_0f(x)$$ and both sides vanish when $$f(x)=0$$.

2. We can restrict $$f(x)\gt0$$ and only consider $$f(x)=0$$ as a limiting case in the closure.

3. Note that for $$f(x)\approx0$$, $$-(1+\log(f(x)))$$ is huge. That means that to maximize $$(2)$$, any place where $$f(x)\approx0$$ we want $$\delta f(x)$$ to be positive, so that $$-\int_a^b(1+\log(f(x)))\delta f(x)\,\mathrm{d}x$$ increases. That is, no maximizing function will have $$f(x)=0$$.

• Welcome to CV. Because most of the readers of this thread will be unfamiliar with the conventions of the Calculus of Variations, it would help to (a) explain that "$\delta f$" is a variation in $f$ and not just some constant times $f$ and (b) show how you went from (2) to (3) and how you applied orthogonality to obtain (5). Please note, too, that the family of continuous distributions includes functions $f$ that can have values of zero. Thus, to make sense of integrating $1+\log(f(x))$ and invoke orthogonality you need to work in the extended reals.
– whuber
Mar 26, 2020 at 19:59
• @whuber: thanks for the comment. I have hopefully covered these issues in my edit. I prefer to avoid the extended reals, so I offer a few other workarounds. Mar 27, 2020 at 8:54
• Well done and thank you for the explanations! (+1)
– whuber
Mar 27, 2020 at 14:24

An intuitive explanation:

If we put more probability mass into one event of a random variable, we will have to take away some from other events. The one will have less information content and more weight, the others more information content and less weight. Therefore the entropy being the expected information content will go down since the event with lower information content will be weighted more.

As an extreme case imagine one event getting probability of almost one, therefore the other events will have a combined probability of almost zero and the entropy will be very low.