You are here

stage based modelling

Hello,

I am new to modelling and simile, and dont have a fantastic
mathematical ability.

I have a population of logs in Decay class 1, and i know that
members move from decay class 1 to 2 between 20 and 50 years, with the mean at 35 years.

How can i capture this pattern around the mean?
can i draw a normal distribution graph of the mean and the variance and incorporate this into the flow out of the decay class 1 compartment?

i can send further info and model if needed

cheers,

LS

Forums: 

Hiya,

If you are aggregating all your logs into a compartment representing the decay class, you cannot ensure that they all move on between 20 and 50 years after being added. The usual approximation is to move a fraction of the contents of one compartment to the next each year.

For a more accurate representation than this you could disaggregate your logs into a population. You would then calculate the time to decay of each log as it was added to class I, either by a sketch graph or by the gaussian() function, and have it change class after this time.

Hope this helps
--Jasper

Thanks

I finally realised that gaussian was fancy speak for that distribution

Cheers

LS

Hello,

thanks for your help with the previous question.

I was modelling the change of decay class based on age of the log

if age>gaussian(index(1),35,10) then Decay class 1 moves to Decay class 2 and so on for each decay class

that is good for the first class as i am assuming that the logs enter decay class
1 at the same time.

but what i want to do is specify the time that the log entered the decay class 2 compartment and specify a 'residence time' of the log in that class based on when it arrived in the compartment.

eg a log may enter the decay class 2 at a time of 20 and from then it may stay in decay class 2 for an average of 50 years with a standard deviation of around 15 years.

i know there is a simple answer to my question, i just havent found it yet.

cheers

LS

Hi LS,

The function you are looking for is init_time (http://www.simulistics.com/help/equations/functions/init_time.htm) it returns the time at which when a model component first came into existence.

You can then model the residence time in Decay Class 2 in the same way you modelled the age at which a log should be removed from Decay Class 1.

Jonathan.

...alternatively, if you just have a variable to say which decay class a log is in, you can record the time at which it changes with a function such as:

if class != last(class) then time() else prev(1)

This will be equal to the time at the point at which class changes, and keep its previous value when class does not change

--Jasper

D Thankyou thankyou thankyou,

that solved my dilemma for the moment.

out of curiosity, where would i have found this formula/solution
(if class != last(class) then time() else prev(1))
on simulistics.com?

is it possible to give small worked examples of this kind in a
seperate 'solutions' area of the site that one could look at and incorporate into their own model. rather than looking at worked models which are more confusing.
perhaps this is too simplistic.

Regards

LS

Well there are entries on the time() function and if...then... construct in our Tips & Tricks section of the website, but it's not very easy to search for the answer to any given problem. We need to add some sort of keyword-search to it -- also some means of attaching keywords to topics in this forum might be useful. What happens if I put something in the Subject field?

There is the "Tricks and Tips" section of the web site http://www.simulistics.com. (Does this fit with your idea of a "solutions" section? Let us know of any suggestions you have for improving it.) Of course, that section is rather small but we do plan to expand it using material from support emails and this forum. So, if the "Trick and tips" section does not answer your problem the best thing is to post on this forum where we will do our best to help. Material on the forum is then available for others to see and incorporation by us into the "Tricks and tips".

Jonathan
Simulistics