Appendix A: Generating Random Numbers from a Distribution: Difference between revisions

From ReliaWiki
Jump to navigation Jump to search
(25 intermediate revisions by 7 users not shown)
Line 1: Line 1:
{{Template:bsbook|A3}}
{{Template:bsbook SUB|Appendix A|Generating Random Numbers from a Distribution}}
 
Simulation involves generating random numbers that belong to a specific distribution. We will illustrate this methodology using the Weibull distribution.  
Simulation involves generating random numbers that belong to a specific distribution. We will illustrate this methodology using the Weibull distribution.  


=Generating Random Times from a Weibull Distribution=
=Generating Random Times from a Weibull Distribution=


The <math>cdf</math>  of the 2-parameter Weibull distribution is given by,
The ''cdf'' of the 2-parameter Weibull distribution is given by:
<br>
 
 
::<math>F(T)=1-{{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}.</math>
<br>


::<math>F(T)=1-{{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}\,\!</math>


The Weibull reliability function is given by,
The Weibull reliability function is given by:
<br>
 


::<math>\begin{align}
::<math>\begin{align}
  R(T)= & 1-F(t) \\  
  R(T)= & 1-F(t) \\  
  = & {{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}.  
  = & {{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}   
\end{align}</math>
\end{align}\,\!</math>


<br>
To generate a random time from a Weibull distribution, with a given  <math>\eta \,\!</math>  and  <math>\beta \,\!</math>,  a uniform random number from 0 to 1,  <math>{{U}_{R}}[0,1]\,\!</math> , is first obtained.  The random time from a weibull distribution is then obtained from:


To generate a random time from a Weibull distribution, with a given  <math>\eta </math>  and  <math>\beta </math>  a uniform random number from 0 to 1,  <math>{{U}_{R}}[0,1]</math> , is first obtained.  The random time from a weibull distribution is then obtained from:
::<math>{{T}_{R}}=\eta \cdot {{\left\{ -\ln \left[ {{U}_{R}}[0,1] \right] \right\}}^{\tfrac{1}{\beta }}}\,\!</math>
 
<br>
 
::<math>{{T}_{R}}=\eta \cdot {{\left\{ -\ln \left[ {{U}_{R}}[0,1] \right] \right\}}^{\tfrac{1}{\beta }}}.</math>
 
<br>


==Conditional==
==Conditional==
The Weibull conditional reliability function is given by:


The Weibull conditional reliability function is given by,
::<math>R(t|T)=\frac{R(T+t)}{R(T)}=\frac{{{e}^{-{{\left( \tfrac{T+t}{\eta } \right)}^{\beta }}}}}{{{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}}\,\!</math>


<br>
The random time would be the solution for <math>t\,\!</math> for <math>R(t|T)={{U}_{R}}[0,1]\,\!</math>.
 
::<math>R(T,t)=\frac{R(T+t)}{R(T)}=\frac{{{e}^{-{{\left( \tfrac{T+t}{\eta } \right)}^{\beta }}}}}{{{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}},</math>
 
<br>
 
or,
<br>
 
<br>


=BlockSim's Random Number Generator (RNG)=
=BlockSim's Random Number Generator (RNG)=


Internally ReliaSoft's BlockSim uses an algorithm based on L'Ecuyer's [RefX] random number generator with a post Bays-Durham shuffle.  The RNG's period is aproximately <math>{{10}^{18}}</math> , which is more than sufficient since this number is larger than the maximum number of simulations allowed by BlockSim. The RNG passes all currently known statistical tests, within the limits the machine's precion, and for a number of calls (simulation runs) less than the period. If no seed is provided the algorithm uses the machines clock to initialize the RNG.
Internally, ReliaSoft's BlockSim uses an algorithm based on L'Ecuyer's [RefX] random number generator with a post Bays-Durham shuffle.  The RNG's period is approximately 10^18. The RNG passes all currently known statistical tests, within the limits the machine's precision, and for a number of calls (simulation runs) less than the period. If no seed is provided the algorithm uses the machines clock to initialize the RNG.
 
 


=References=
=References=
*> L'Ecuyer, P., 1988, Communications of the ACM, vol. 31, pp.724-774
#L'Ecuyer, P., 1988, Communications of the ACM, vol. 31, pp.724-774
<br>
#L'Ecuyer, P., 2001, Proceedings of the 2001 Winter Simulation Conference, pp.95-105
<br>
#William H., Teukolsky, Saul A., Vetterling, William T., Flannery, Brian R., Numerical Recipes in C: The Art of Scientific Computing, Second Edition, Cambridge University Press, 1988.
L'Ecuyer, P., 2001, Proceedings of the 2001 Winter Simulation Conference, pp.95-105
#Peters, Edgar E., Fractal Market Analysis: Applying Chaos Theory to Investment & Economics, John Wiley & Sons, 1994.
<br>
#Knuth, Donald E., The Art of Computer Programming: Volume 2 - Seminumerical Algorithms, Third Edition, Addison-Wesley, 1998.
<br>
Press, William H., Teukolsky, Saul A., Vetterling, William T., Flannery, Brian R., Numerical
Recipes in C: The Art of Scientific Computing, Second Edition, Cambridge University Press, 1988.
<br>
<br>
Peters, Edgar E., Fractal Market Analysis: Applying Chaos Theory to Investment & Economics, John Wiley & Sons, 1994.
<br>
<br>
Knuth, Donald E., The Art of Computer Programming: Volume 2 - Seminumerical Algorithms, Third Edition, Addison-Wesley, 1998.
<br>

Revision as of 21:41, 8 February 2017

Template:Bsbook SUB Simulation involves generating random numbers that belong to a specific distribution. We will illustrate this methodology using the Weibull distribution.

Generating Random Times from a Weibull Distribution

The cdf of the 2-parameter Weibull distribution is given by:

[math]\displaystyle{ F(T)=1-{{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}\,\! }[/math]

The Weibull reliability function is given by:

[math]\displaystyle{ \begin{align} R(T)= & 1-F(t) \\ = & {{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}} \end{align}\,\! }[/math]

To generate a random time from a Weibull distribution, with a given [math]\displaystyle{ \eta \,\! }[/math] and [math]\displaystyle{ \beta \,\! }[/math], a uniform random number from 0 to 1, [math]\displaystyle{ {{U}_{R}}[0,1]\,\! }[/math] , is first obtained. The random time from a weibull distribution is then obtained from:

[math]\displaystyle{ {{T}_{R}}=\eta \cdot {{\left\{ -\ln \left[ {{U}_{R}}[0,1] \right] \right\}}^{\tfrac{1}{\beta }}}\,\! }[/math]

Conditional

The Weibull conditional reliability function is given by:

[math]\displaystyle{ R(t|T)=\frac{R(T+t)}{R(T)}=\frac{{{e}^{-{{\left( \tfrac{T+t}{\eta } \right)}^{\beta }}}}}{{{e}^{-{{\left( \tfrac{T}{\eta } \right)}^{\beta }}}}}\,\! }[/math]

The random time would be the solution for [math]\displaystyle{ t\,\! }[/math] for [math]\displaystyle{ R(t|T)={{U}_{R}}[0,1]\,\! }[/math].

BlockSim's Random Number Generator (RNG)

Internally, ReliaSoft's BlockSim uses an algorithm based on L'Ecuyer's [RefX] random number generator with a post Bays-Durham shuffle. The RNG's period is approximately 10^18. The RNG passes all currently known statistical tests, within the limits the machine's precision, and for a number of calls (simulation runs) less than the period. If no seed is provided the algorithm uses the machines clock to initialize the RNG.

References

  1. L'Ecuyer, P., 1988, Communications of the ACM, vol. 31, pp.724-774
  2. L'Ecuyer, P., 2001, Proceedings of the 2001 Winter Simulation Conference, pp.95-105
  3. William H., Teukolsky, Saul A., Vetterling, William T., Flannery, Brian R., Numerical Recipes in C: The Art of Scientific Computing, Second Edition, Cambridge University Press, 1988.
  4. Peters, Edgar E., Fractal Market Analysis: Applying Chaos Theory to Investment & Economics, John Wiley & Sons, 1994.
  5. Knuth, Donald E., The Art of Computer Programming: Volume 2 - Seminumerical Algorithms, Third Edition, Addison-Wesley, 1998.