Another try at teaching simulations

[Rayman Mohamed]

Hi all,

I received some good responses to my last question on this list, so I am 
going to try my luck again on this one.  I am also sure that related forms 
of this question have been aired here before.  (In fact I checked my 
swarm-modeling mailbox  that's a few months old and found related issues 
but not quite the same thing.)

Here is the context -- please bear with me, I think that it is 
important.  As part of my TA responsibilities this semester, I am working 
with a professor updating a course for next semester called something like 
"Techniques for Policy Analysis."  The course is offered by the Dept of 
City and Regional Planning but 90% of the students are from Public 
Affairs.  The course had been taught for many years and is incredibly 
fast-paced.  One of the principal techniques taught is simulation modeling, 
which for City and Regional Planners goes back to the mid 1960s.  The book 
uses an outdated, out-of-print manuscript by a guy by the name of Forrester 
from MIT prepared circa 1969.  He, apparently, thought of urban systems as 
stocks and flows.  (This is unlike more formal urban simulations based on 
economic theories.)  The professor uses this same "thinking" to deal with 
public policy scenarios, and if I may say so, I would argue that he was 
ahead of his time when he started doing this sometime in the 1970s (I 
guess).  History also indicates that the professor was quick to adopt 
spreadsheets as a simulation tool (see comment on this later on.)  Typical 
examples of case studies in class might be: modeling arrivals at a refuge 
camp, emergency room situations, teacher supply in a developing country, 
etc.  The examples have no spatial context.

About the students -- most are extremely hard working but lack math 
skills.  Their undergraduate training is typically in political science, 
anthropology, international studies, agriculture, etc., or something of 
that nature.  Most do not understand concepts of probability.

However, it is time to "update" the course again and here are the principal 
problems that the professor and I have with the simulation component of the 
course.  First, there is no good textbook we know of to recommend to 
students.  Second, no stochastic concepts are built into the simulations 
(if you can call it that without any stochasticity).  Third, agents are 
homogenous -- there is no individual behavior.  Fourth, there is no 
emergent behavior -- everything is forced into the simulation.    (If 
anybody can think of more problems do not hesitate to say so, but perhaps 
in a separate email.)

Of the problems listed above, we believe that the first three are the most 
important and if we were to narrow it down further we would say that the 
biggest problem is the textbook.  So, after all that I guess I am asking if 
anybody can recommend a good text for the situation I described above.  An 
ideal text from our perspective would contain an introduction to 
probability, the concept of different agents doing their own thing, and 
dealing with all this using Excel as the tool of choice.   I guess that the 
title of my ideal text would be: "Simulating Modeling in Excel: Public 
Policy Applications."

I can imagine some frowns as you read this email -- Excel???  I guess that 
I need to justify this a little further.  First, neither the professor nor 
I would feel comfortable teaching these students to write computer 
code.  Besides being complete novices ourselves, we do not have the time as 
this part of the course in taught in a month.  Second though, and more 
importantly I would argue, I do believe that Excel can meet the needs of 
what we are trying to do as long as one figures out how to handle 
probability in it, and one is willing to keep the situation being modeled 
simple -- may have to sacrifice some of the "individualness."  At the same 
time, we are open to suggestions about other software packages you thing 
might meet our needs -- given the comments in this paragraph Swarm and the 
like are out of the question.  On the other hand, the professor is set 
against the use of "black box" software.  In fact, that is one reason why 
he likes Excel -- it forces students to think as their simulations have to 
expressed through "coding" in Excel, however rudimentary it might be.

Ok, I hope that I have made the situation clear enough.  And, I hope that a 
good text exists out there that one of you knows of.  (If not, I do believe 
strongly that the market exists for such a text.)  Thanks for you insights,



Rayman



                 ==================================
  Swarm-Modelling is for discussion of Simulation and Modelling techniques
  esp. using Swarm.  For list administration needs (esp. [un]subscribing),
  please send a message to <address@hidden> with "help" in the
  body of the message.
                 ==================================