looking for help--why the curves end up in different extremu

[lookingfor17]

I am a student in China. And I have learned a little about SD and Vensim.
Now I am working for my paper and I want to build a SD model. After I ran my simulation model, I found something wrong.
When I changed a rate, the S-shape curve seemed much higher or lower than the baserun. But these S-shape curves were supposed to be end up at a same extremum. I tried my best to fix it. I spent two days and still cannot find the solution.
Nobody around me studies SD and I have nobody to ask.
Will anybody here can tell me why? I have to hand in my paper as soon as possible.
Thank you very much!!!

[Edited on 4-16-2010 by lookingfor17]

[Edited on 4-16-2010 by lookingfor17]

[Edited on 4-16-2010 by lookingfor17]

[Edited on 4-21-2010 by lookingfor17]

[Administrator]

We cannot really offer any advice unless you can upload the model you need help with.

Tony.

[lookingfor17]

I have uploaded my model. I am sorry that I wrote it in Chinese. This may cause some trouble when you read it.
If you think it is better to translate it into English, tell me and I will do it.

My problem is:
I need to research the curve of 培养水平. When I change some rate of the model, the curves of 培养水平was supposed to be end up at a same extremum. But my model cannot do this.

Thank you very much!

Joy

[bob@vensim.com]

From the equations there is no reason to expect the level to end at the same value. For this to happen there needs to be a negative feedback look that ends basically comparing the level to the final value and adjusts either the inflow or the outflow to make that happen. A couple of hints:

Be sure to add units of measure to every variable and do units checking.

Download the newest Vensim PLE from http://www.vensim.com and convert the model to unicode - you will get much better results (and there is also a Chinese menu system available).

[lookingfor17]

Thank you for your hints. I will take your advice.

But I still cannnot understand how to make a "negative feedback look that ends basically comparing the level to the final value ". I have linked a lookup which slowly goes down. I have thought that it could make this feedback a negative one. Am I wrong?

Or it is only the problem to adjust the value of constants? I have tried thoustands of times and cannnot make it.

I am not a skillful model-maker. Can you tell me which variable or constant should I fix to make that happen?
By the way, this model comes from Peter Senge`s "the limit of growth" feedback. The levels of final value should be the same.

Thank you!!!

Joy

[bob@vensim.com]

A common limits to growth approach is

deaths = population/average lifetime
average lifetime = reference lifetime * effect of crowding on death
effect of crowding on deaths = effect of crowding on deaths f(population/reference population)

where the function is an increasing function. While this is a limits to growth formulation, but it does not mean the same population is always achieved. In this case the stronger the positive feedback loop driving growth the higher the resulting steady state population. This is actually typical of such benign limits to growth formulations.

[lookingfor17]

Ahh~ I see!!!

I shouldn`t have such a hypothesis! And the result of model running is normal!

Thank you very very very much!!!
I really appreciate your help!
Now I can finish my paper.

Joy