Triangular distribution (from http://www.math.wm.edu/˜leemis/chart/UDR/UDR.

html)
The shorthand X ∼ triangular(a, m, b) is used to indicate that the random variable X has the tri-
angular distribution with parameters a, m and b. A triangular random variable X has probability
density function 
 2(x−a) a<x<m
(b−a)(m−a)
f (x) = 2(b−x)

(b−a)(b−m) m ≤ x < b.
The triangular distribution can be used as an approximate model when there are no data values.
An expert familiar with the population specifies a minium value a, a most likely value m, and a
maximum value b. The probability density function is illustrated below.

f (x)

x
a m b

The cumulative distribution function on the support of X is


 (x−a)2 a<x<m
(b−a)(m−a)
F(x) = P(X ≤ x) =
 1 − (b−x)2 m ≤ x < b.
(b−a)(b−m)

The survivor function on the support of X is


 1 − (x−a)2 a<x<m
(b−a)(m−a)
S(x) = P(X ≥ x) =
 (b−x)2 m ≤ x < b.
(b−a)(b−m)

The hazard function on the support of X is

(
2(a−x)
f (x) ab−mb+ma+x 2 −2ax a<x<m
h(x) = = 2
S(x) m ≤ x < b.
b−x

The moment generating function of X is

2(−meta + beta + betm + aetm − aetb + metb )
M(t) = E etX = −
 
t > 0.
(a − m)(a − b)(m − b)t 2

1
The characteristic function of X is
  2(−meita + beita + beitm + aeitm − aeitb + meitb )
φ(t) = E eitX = t > 0.
(a − m)(a − b)(m − b)t 2

The population mean, variance, skewness, and kurtosis of X are

a+m+b a2 + m2 + b2 − ab − am − mb
E[X] = V [X] =
3 18
" 3 # √ " 4 #
X −µ 2(a + b − 2m)(2a − b − m)(a − 2b + m) X −µ 12
E = E = .
σ 2 2 2
5(a + b + m − ab − am − bm) 3/2 σ 5
APPL verification: The APPL statements

X := TriangularRV(a, m, b);
CDF(X);
SF(X);
HF(X);
Mean(X);
Variance(X);
Skewness(X);
Kurtosis(X);
MGF(X);

verify the cumulative distribution function, survivor function, hazard function, population mean,
variance, skewness, kurtosis, and moment generating function.