Sunday, November 24, 2013

System Dynamics as a tool for modelling Complex systems

Over 40 years ago, a visionary character proposed a way of modelling systems based on the control systems theory that was being used at the time in the electrical engineering area, and which proposed the understanding of a system in terms of the different interactions between its components and the stock and flow of relevant quantities. The idea at first may seem simple, but it has deep and very profound consequences.
Let us do a simple mind exercise. Think about a bathtub. This bathtub is originally empty, but might be filled by opening the water tap above it, and closing the drain. In this way, the bathtub may be filled until a certain depth. What would happen if we open the drain? Well it depends. If the water escaping through thee drain is greater than the water coming in through the water-tap, then the water level inside the drain will slowly being to decrease. The contrary will happen if the water coming in is greater than the water leaving the bathtub through the drain. Now imagine the "Water" inside the Bathtub was instead the Greenhouse gasses in the earth's atmosphere. The dynamics of accumulation, as well as delay in closing or opening the "taps" and "drain" for the greenhouse gasses, are in this way much more understandable.

This example, from which a parallel can be drawn to our own experiences, is a very simple conceptualization of the stock, low and delay concept central to the System Dynamics framework. The Bathtub can be considered as the stock in this system, and the water coming in through the tap or leaving the bathtub through the sink can be considered to be the flows.
It results that, if thought out carefully, many problems can be represented in these terms, as concepts that can be viewed as stocks include available forests, wildlife diversity, cars in a specific street at a given time, or the experience in building houses at a particular construction company, for instance. The rates at which these quantities change can have a wide variety of influences in the way the complete system behaves over time, can expose a causal relationships that can be quite complex, and a resulting behavior that is highly non-linear.
There is extensive literature on the application of these techniques in industrial settings, yet from the information I have been able to gather so far, I am for the time being convinced that at MIT I have come across a discipline which is not getting nearly as much attention as I believe it must, and I am having a hard time figuring out why . I have spoken with leading practitioners and direct disciples of the inventor of this approach, Jay Forrester, who have been able to provide their particular viewpoint, but this matter is something in which I believe much is to be written.

No comments:

Post a Comment