|
Simulation: the
inversion transform method
Simulates one value from a selected probability distribution, for a given
or random value, showing the inversion transform method. |
|
|
| Distribution |
Gauss
triangular [in
(μ ± a)] |
Choose distribution: Gaussian or (symmetrical)
triangular. |
| μ, dispersion |
|
Mean and (for G.) σ or (for tri.) a. |
| rand, x0 |
|
Random number and initial guess for
the numerical calculation (see below). |
| .Seed |
|
(Non-negative integer) ‘Seed’ for simulation (see
below). |
One simulated value, a non-uniform (pseudo-)random number,
is calculated from the distribution chosen. The inversion transform
method is applied (using the Newton-Raphson iterative procedure).
For null 'seed', 'rand' is ignored.
Suggested initial guess: μ. For the triangular, the
initial guess must be in (μ ± a).
For rand = 0.99865 and Gauss, it will be
x = μ + 3 σ. (Why ?) |
|
| References: |
Plate sim04a13 |
|
• Wagner, Harvey M.,
1972, “Principles of Operations Research”,
Prentice-Hall International, Inc., London, UK
(2.nd ed.). |
