Continuous uniform distribution (source code)

= Continuous uniform distribution
{title2=$U(a,b)$}
{wiki}

= Uniform distribution
{synonym}

= Uniform density
{synonym}

= Uniformly distributed
{synonym}

A <continuous uniform distribution> on a finite <interval> with endpoints $a<b$ has <probability density function>
$$
f(x)=\frac1{b-a}\mathbf1_{\{a<x<b\}}.
$$
Choosing open or closed endpoints does not change this <probability distribution>, because single points have zero <probability>. Its <cumulative distribution function> is zero below $a$, $(x-a)/(b-a)$ between the endpoints, and one above $b$. Consequently a subinterval $[c,d]\subseteq[a,b]$ has <probability> $(d-c)/(b-a)$.

Integration gives <expected value> $(a+b)/2$ and <variance> $(b-a)^2/12$. If $U$ is uniform on $(0,1)$, then $a+(b-a)U$ is uniform on $(a,b)$, a basic input transformation for <Monte Carlo method> and <inverse transform sampling>. For example, on $(2,5)$ the <expected value> is $7/2$, the <variance> is $3/4$, and $\mathbb P(3<X<4)=1/3$. A <uniform distribution> is relative to its coordinate or reference <measure>: taking its <logarithm> generally does not preserve uniformity, as the <log-uniform distribution> illustrates.