Subject: help in pendulum model

Posted by Paul Newton on 9/30/2008
In Reply To:help in pendulum model Posted by Karim Chichakly on 9/30/2008

Message:

It seems to me that Section 13.3.1 in Sterman's Business Dynamics textbook entitled "All Outflows Require First-Order Control" is essential reading for those intersested in this discussion thread.

From the first three sentences in that section: "Real stocks such as inventories, personnel, cash, and other resources cannot become negative. You must formulate the rates in your models so that these stocks remain nonnegative even under extreme conditions. Doing so requires all outflows to have first-order control. "

The default settings in Stella and ithink automatically, and hidden from the unsuspecting modeler, provide this first-order control.

The problem with this is that a modeler can write an outflow equation (or a net inflow equation that can become negative) that is faulty in that it would cause stocks to go negative when they shouldn't. But when testing her model under the default non-negative stocks setting of Stella/ithink, the modeler would never observe this faulty behavior. So she would be lulled into thinking that she had a good equation formulation, when in fact, the formulation is not representative of how the real world works.

Hope that makes sense...

I've never fully understood why Barry included this default setting in Stella/ithink, but I suspect it had to do with focusing new modelers on the major structures that create the dynamics they're concerned about. However, my sense is that this has had the unintended consequence of a continued lack of appreciation of the value of testing for negative stocks, even by quite experienced modelers, thus limiting modelers' ability to discover errors in their thinking about structure.

Follow Ups:

help in pendulum model - Bill Braun 9/30/2008