Skip to main content

implementation of a lognormal distribution


I would like to implement a lognormal distribution in AIMMS. I have the parameters mu and sigma as defined on Wikipedia. https://en.wikipedia.org/wiki/Log-normal_distribution. How do I implement these into AIMMS to get the same pdf. I tried to understand the formula in the AIMMS_ref.pdf documentation but implementation of the transformation remains unclear.

3 replies

deannezhang
AIMMS Champ
Forum|alt.badge.img+3
  • AIMMS Champ
  • 60 replies
  • August 19, 2019
As what is described in the AIMMS function reference, the function LogNormal( Shape,Lowerbound, Scale ) takes three input parameters, where

Shape = sigma/mu,
Lowerbound = 0
Scale = mu

Jess Es
Forum|alt.badge.img+6
  • Grand Master
  • 120 replies
  • August 23, 2019
Hi @Jan Willem ter Weeme - did you find the information you needed? Do you mind to share how it worked out? (in case others have a similar issue)

Hi @deannezhang, @deannezhang. Thanks for getting in contact. In the end it turned out the implementation needed to be a bit different. The request was to implement:


With a mean value of 0.057 and a sigma of 0.1. In the end I implemented simply
lognormal(sigma,0,mean)
which in Minitab yielded exactly the distribution I needed (see mean and scale)


So the division of the shape by mu was not needed.
Remains that to my opinion the explanation in the manual is quite hard to follow.

Reply