Vensim for Dummies?

[Odlanierzenitram]

Hello everyone. I have registered to learn a bit about Vensim. I must warn you all that I know very little about systems and am lousy at maths, but since this doesnt seem to be a need with Vensim I am in.

Q1.-Can anyone please tip me on how to use percetages in Vensim equation variables? So many butttons confuse me. I see no Percentage in the "UNITS" menu, but persons and dollars an years. etc.

Q2. Can anyone please point out the difference between quantitative and qualitative models in Vensim? Can both be built?

Thanks to you all and please send tips to Odlanierzenitram at Yahoo dot com.

[Administrator]

I can answer Q1. Percentages are dimensionless, so in the units box, just add DMNL or percent (or spell it out in full if you need to).

Any parameter can be any dimension. When you run the units check, Vensim makes sure the units on one side of the equation match the other. It just gives a lot more confidence that a model is calculating correctly (eg, you are not entering delays in "months" and using them as "days").

[Odlanierzenitram]

Thanks, Adm. Sorry if I sound real dumb, but what are the two sides of the equation? Units has only one window. THere are other three indicatinm "minimum value maximum value and increment. I understand that these are the starting ammount and the top of the graph.

I need for instance, tohave % of unemployment in which 3% is the topmost positive ammount: 0% through 3% will be negative 3% to 4% should be positive and 4% through 9% will be also negative.

But I don't know how to do it.

Thanks anyway.

[Administrator]

Equations must always balance on both sides.

Eg, think of the equation "distance = time_doing_speed * speed"

distance = km
speed = km/hour
time_doing_speed = hour

Put the units on both sides gives
km = hour * km/hour

On the right hand side of the equation, the "hour" cancels out. In this case, the equation is dimensionally correct as km = km. Vensim checks all of these.

[tomfid]

I typically add a units conversion variable to my model, so that I don't mix up % and fractions. For example:

"100 percent" == 100 ~percent ~this is the conversion factor for percentages, with units of percent; note the == makes it an unchangeable constant
unemployment rate = 3.5 ~percent ~a typical constant, defined as a percentage
unemployed people = working population *unemployment rate/"100 percent" ~ people ~ a calculation where the "100 percent" constant is used to convert % to a fraction

For my own stuff, I almost never use percent; I just use fractions directly (i.e. unemployment rate = 0.035 ). I only use percent in situations where a user might not be used to thinking that way.

[LAUJJL]

Hi Tom

All percentage constants are the result of the division of an unit by the same unit. So if one uses the dimension unit/unit
one sees that it is a percentage and one sees too what kind of percentage it is, which is more informative to my opinion.
See attached the simplistic model for the unemployed people.
Regards.
JJ

[Lee Jones]

JJ, see the model attached, your model included unemployment rate as a fraction. In the attached model, a conversion variable is added to allow me to add "unemployment rate %", converting the percentage value into a fraction for use in the equation.

It's personal choice as to whether to use a fraction or %.

[LAUJJL]

Hi Lee

I have understood Tom's explanations that you reproduce in the percentage_2 model.
but I do not see the advantage of using the percent this way.
Why not simply define the unit of the rate as:
percent/percent that will show that we are dealing with a percentage and will avoid defining the constant 100% percentage equal to 100 percents.
See the model joined that simplifies your model while being always as explicit.
On the contrary my model percentage is much better to my point of view because it tells that the percentage is the fraction of people by people, that is dimensionless as the fraction percent/percent withe the advantage of showing that it is a fraction of people.
Regards.
JJ

[Lee Jones]

Hi JJ,

I suppose it's a matter of semantics; percent (per hundred) to me would mean that 0.1 = 0.1 percent (0.001 as a fraction). I would expect to see the value "10" if the input was 10% and not the value 0.1.

Lee

[Odlanierzenitram]

I believe I need to look for answers elsewhere. Not being a system master I got dizzy trying to follow the thread, since I only see specialists talk. I still have no clue about what to do with the Vensim menus to achieve what I want.

Is there any begginer's forum somewhere?


I am sorry., but thanks you all anuway.

[bob@vensim.com]

The difficulty is not specialist talk but getting an understanding of what you are trying to achieve. If you are simply trying to graph data, you might want to use Excel as this will likely be easier for you to work with. If you are trying to build a simulation model you should work through Chapter 6 of the User's Guide that will take you through this process.

[LAUJJL]

Hi Odlanierzenitram (not esasy to spell or write!)

SD is not easy to learn and while it requires no calculus (differential or integral calculus) it is to my opinion mathematically speaking as difficult.
SD needs to be studied quietly and takes time.
The best thing is first to study the Vensim user guide, chapter by chapter. A chapter will be understood only if you are able to build by yourself the models exposed in each chapter without copying the book.
You can avoid chapters on subscripts that is a bit difficult to understand and keep it for later studying.
This will take several months, if you do it seriously.
After that I would suggest reading a book like Coyle's 'Sytem dynamics a practical approach' and translate all the examples given in Cosmic in Vensim. If you have difficulties in doing that I can help you as I have translated most of the models my self.
This book will give you a complete methodoly of building models, the only complete methodology I know. A third of the book deals with influence qualitative diagrams which to my opinion is critical to build useful models.
Coyle is too an exceptionnal teacher and all his books are interesting to study.
Studying seriously this book will take you some complementary months.
That is to my opinion enough to start building by yourself useful models.
You will probably too use this forum much better as as you could see it, you will always find somebody to help you.
Before engaging into this effort, I would too verify that you really need to study SD. What is your situation, are the problems you are daling with suceptible to be solved with help of SD?
If you find difficult to study alone, you might too consider following distant web courses. There are two that use both Vensim. The WPI distant courses and one sponsored by the Unesco and taught too in English. You can find references of both courses on the web site of the system dynamic associatiion.
www.systemdynamics.org
Regards.
JJ

[Edited on 8-10-2010 by LAUJJL]

[Odlanierzenitram]

Hi to everyone, thanks for all your help.

I am trying to develop some model for our City's Sustentability Management. Using a Causal Diagram, I have chosen four main variables to feed it, namely. Enviironmental Sustainability, Economic Autonomy and Social Stability, by picking the three standard Goals of the Millenium and many other sources in environment ans sustainability. In adition I have chosen Jack Harich's (www.twink.org) Resistance to Change as the fourth variable, being that this one tends to show up with untold force in Caracas these times (we all now why).

I have also pinpointed at least four variables (to make it conveniently simple for a demo), to each one of the 4 mentioned leading ones. As you can imagine, in order to set equations for all these can become a nightmare, so I decided to use percentages. % of Sustainability, % of Resistance to Change, % of Economic Autonomy and so on.

Of course, the leading variables have been arbitrarily valued at 40% Resistance to Change, and 20% each one of the others. This means that when RtC is "zero" sustainability will be 100%. The more RtC the less sustainability. In turn the more Social Stability the mosre sustainablethe City will be, and so on and so forth. Social and scientifical studies will set truer values to each.

I simply need to set equations to each one of the subvariables conforming each one of the leading variables. That is, I have 'Solid Waste Management', 'Water Quality' (in%), 'Visual Pollution' (in %) and 'Green Areas in % (sqm/inhabitant)' as the 4 subvariables that compose "Environmental Sustainability", one of the four main variables mentioned.

Obviously these can be a dozen, or even three dozen different variables. But I've chosen only four each as to be able to make a demo to the authorities, as we all know that Govt. officer wont ever have time to listen for hours on end to someone explaining a model with 60 or 70 variables.

Delay varies in each case. E. g, in RtC it is zero, but in Environmental Education it ranges from 8 to 20 Quartyers depending on the kind of educational method we may use.

I've found that the learning curve for Vensim is quite steep for non conosseurs of SD. Therefore my question now would be:

What equations may I use in these 16 subvariables and in the leading 4 determinants as to be ableto visualize a graphic of Urban Sustainability and pull sliides here and there to demonstrate the use of such tool to make better decisions?



Thanks a lot to you all.

[LAUJJL]

Hi


What did you take the Millennium goals from?
Is it in relation with the Millennium Institute in Bergen that has taught a course in Vensim this year in Bergen?
Regards.
JJ

[Odlanierzenitram]

No, mate. These 3 base lines (Economy, Environment and Social) are the three poles of any sustainable development as for most international agreements worldwide. Every project sohould comply with these three actgion lines and provide a balance between them to consider it "sustainable". I simply added a 4th, (Resistance to Change) being that it has become such an important issue at home recently.

For your info, this is my first experience with Vensim, It is not as easy as it seems in YouTube!

I think that unless you get someone to instruct you it will take months before you can accomplish a simple model with 16 variables.

Thanks and cheers.

[tomfid]

Don't feel bad. I think what you're experiencing is the huge gulf between a rich intuitive picture of a problem, and a formalized model. The sustainability concepts you're working with are very multifaceted and therefore complex to model.

Fortunately, people tend to have a very limited understanding of such dynamics, so you can make a worthwhile contribution with even a very simple model. For example, people may not have thought about the stock nature of resources, so pointing out simple dynamics of accumulation may be helpful. For example, if you cut fewer trees each year, forests may still shrink, unless you match your rate of harvest to regeneration.

You might take a look at something like World Dynamics and Urban Dynamics for inspiration (these are part of the Vensim sample models, or here http://models.metasd.com/2010/03/world-dynamics/ http://models.metasd.com/2010/06/urban-dynamics/ ).

Good luck.