Rayleigh Distribution
The Rayleigh distribution is a continuous probability distribution named after the English Lord
Rayleigh. The distribution is widely used:
In communications theory, to model multiple paths of dense scattered signals reaching a
receiver.
In the physical sciences to model wind speed, wave heights and sound/light radiation.
In engineering, to measure the lifetime of an object, where the lifetime depends on the object’s
age. For example: resistors, transformers, and capacitors in aircraft radar sets.
In medical imaging science, to model noise variance in magnetic resonance imaging.
The Rayleigh distribution is a special case of the Weibull distribution with a scale parameter of
2. When a Rayleigh is set with a shape parameter (σ) of 1, it is equal to a chi square
distribution with 2 degrees of freedom.
The notation X~~ Rayleigh (σ) means that the random variable X has a Rayleigh distribution
with shape parameter σ.
The Rayleigh distribution is the distribution of the magnitude of two-dimensional random vector
whose coordinates are independent, identically distributed, mean 0 normal random variables.
Definition
Suppose that z1 and z2 are independent random variables with standard normal distributions.
The magnitude R z12 z22 of the vector z1 , z2 has the standard Rayleigh distribution.
Note that z1 , z2 has the standard bivariate normal distribution.
x ~ Rayleigh( ) is used to indicate that the random variable x has the Rayleigh distribution
with parameter σ. A Rayleigh random variable x with parameter σ has probability density
function
x x 2 2
2
e ,x0
f x 2
0 , x 0
As the shape parameter increases, the distribution gets wider.
Rayleigh distribution showing several different shape parameters, σ.
Variance and Mean (Expected Value) of a Rayleigh Distribution
The expected value (the mean) of a Rayleigh distribution is:
How this equation is derived involves solving an integral, using calculus:
The expected value of a probability distribution is:
E(x) = ∫ xf(x)dx.
Substituting in the Rayleigh probability density function, this becomes:
This Wolfram calculator will solve the integral for you, giving the Rayleigh expected value of σ
√(π/2)
The variance of a Rayleigh dist is derived in a similar way, giving the variance formula of:
Var(x) = σ2((4 – π)/2).
x2
The cumulative distribution F x 1 e 2 2
, x0
R elated distributions
R ~ Rayleigh( ) is Rayleigh distributed if R x 2 y 2 , where x ~ N (0, 2 ) and y ~ N (0, 2 )
are independent normal random variables.
- The chi square distribution with 2 degrees of freedom is equivalent to Rayleigh
distribution with σ = 1 i.e. R ~ Rayleigh(1) , then R 2 has a chi squared distribution with
parameter n = 2 d.f
n
- If R ~ Rayleigh( ) , then R
i 1
i
2
has a gamma distribution with parameters n and 2σ2
- The Rice distribution is a generalization of the Rayleigh e.t.c
Applications
An application of the estimation of σ can be found in magnetic resonance imaging (MRI).
As MRI images are recorded as complex images but most often viewed as magnitude
images, the background data is Rayleigh distributed. Hence the above formula can be
used to estimate the noise variance in an MRI image from background data.
The principle of Rayleigh distribution was also employed in the field of nutrition for
linking dietary nutrient levels and human and animal responses. In this way the
parameter σ may be used to calculate nutrient response relationship.
Example
A turbine is to be installed at a site for which the only information is that the average
wind speed is 9.6m/s.
N.B If all one knows about a site is its average wind velocity, v , the best one can do is to
make the assumption that the wind obeyed a Rayleigh distribution, that is , the p.d.f is
v 1 2
f (v) 2 exp (v 2 )
2
Where σ is the mode of the distribution, i.e. the value at which the probability density
function peaks. σ is not the mean value (in this case v ). However there is a relation between
2
the average wind density and the mode of the Rayleigh p.d.f: 2 v 2
v v 2
Thus f (v) 2 exp
2v 4 v
i) How many hours per year will the wind velocity (probably) exceed 12 m/s?
ii) What is the probability of having a wind larger than 11.5m/s and smaller than
12.5m/s.
N.B In case of wind velocities , one can tabulate and plot the probability of having a
wind velocity in a given interval, say v1 v v2 .
To find the probability of having a wind velocity larger than v1 and smaller than v2 , one
v2
forms the integral prob f (v)dv
v1
An informative way of examining a probalistic behavior is to plot the cumulative distribution ,
v
c.d.f. i.e. plot the integral c.d . f (v)
f (v)dv versus the upper limit, v. The c.d.f. answers the
question of how often a given random variable, V , is below a given value , v. By the same
v
token, the complimentary cumulative distribution function c.d.f c.d . f (v) 1
f (v)dv ,
answers how often the variable V is above a given value v.
Solution
i) To answer this question we use the complimentary cumulative Rayleigh distribution.
v 2
c.d . f exp 0.29 where v 12m / s
4 9.6
There is a 29% probability of the wind being faster than or equal to 12m/s. since
there are 8760 hrs in a year, such a wind should blow 0.29 x 8760 = 2568 hrs in a
year.
If the cutting in wind velocity (the velocity at which the turbine starts to generate
electricity) is 3.5 m/s , we can find that the turbine will produce a useful output 90 %
3.5 2
of the time. i.e exp 0.9
4 9.6
ii) P 11.5 v 12.5 cdf (11.5) cdf 12.5 0.324 0.264 0.06
There is a 6 % probability that there be a wind with velocity larger than 11.5 m/s and
smaller than 12.5 m/s