looking to exchange ideas about a specific modeling method

This forum is for discussion of topics related to System Dynamics. You must be logged in to post but anyone who has registered on the System Dynamics web portal may post. Note that there may be a delay of a working day or two after registering before you can post anything.
Post Reply
Jean-Jacques Lauble
Posts: 74
Joined: Fri Aug 23, 2013 3:49 pm

looking to exchange ideas about a specific modeling method

Post by Jean-Jacques Lauble » Tue Feb 14, 2017 4:42 am

Hi everybody

I have been studying SD for 15 years, trying to build useful models, with different methods.

I must confess with very small successes, being able to build credible models but if too complex not being able to build out of them credible policies.

I have recently instead of observing the problem, observed the way I was working and how I could improve it.

I am presently working in a totally different way that I resume shortly and that will change certainly with the time:

Instead of concentrating on the reality, I simplify it first to the maximum imaginable extent, and this extent is much higher than one could imagine. This simplification must start with a top down approach, starting at the top and not at a too much detailed part of the reality. (something that a consultant did for me some years ago).

I then work on this simplified reality, build a model, analyses it, find policies solving the simplified problem and spend the necessary time (which can be much longer that one would imagine) until it is obviously not possible to make it better nor to build a better policy, not to get more insights out of it. Until one is sure that one has reached a plateau in the quality of the model (which will be quite small at the first stages), the quality of the policies, and the overall understanding of the simplified reality.

Of course, the question I ask myself is the following:

Is it possible for a modeller to reach this plateau, without the close participation of people knowing the reality, the decision makers and the people having to implement the policies.

I cannot answer this question, because I have the chance to know the reality, take the decisions and have to implement them. Up to what point is it possible to make these people participate and if not up to what point it is still possible to work correctly?

Of course I keep a trace of all the work done at this stage, so that it is easy to go back to it, later (weeks, months, years). The work must be reproducible easily later, once done.

One this is done, I make this simplified reality, slightly little closer to the reality adding some more definitions or eventually only one and start the whole process again. How high is the step between two stages is a matter of experience. I work presently with extremely small steps to increase the chance of success.

If at one stage, I find it impossible to progress, I consider stepping back and finding a different step or a smaller one until I have find some ways to solve this difficulty. If whatever I do, I stay blocked at this stage, it is simply that I have tried to solve a problem that is or impossible to solve, or impossible to solve by myself, taking into consideration my experience and the tools I use, and it is better to stop at this stage and try finding something else to work on.

The drawbacks of this method is the time needed that can be considerable. It can be cumbersome, laborious and boring. The necessity or not to work with specific people may be a problem.

The expected (my expectations) advantages (still to be proven) are the increased creativity of the policies found, their effectiveness and the quality of understanding of the people concerned (the clients). I have already worked like this for a time and it looks very promising, even if it obliges me to change deeply the way I work.

In fact I am applying Coyle's advice: 'the key to successful modelling is to keep one's understanding of the model and what it says about the problem ahead of its size'.
I apply it replacing 'ahed' by 'far or way ahead'.


I am trying to find people who have made experiences in this direction, preferably with real problems.

Anybody interested by exchanging experiences can answer to that post or join me directly at my e-mail address:

jean-jacques.lauble@estvideo.fr

Best regards.

JJ

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest