On Transdomain Modeling

I received this interesting article URL https://archive.nytimes.com/opinionator.blogs.nytimes.com/2009/06/02/guest-column-like-water-for-money/ from a friend few months back. It talks about a water/glass/plumbing based model that is designed to mimic the world of economics. :-)

Creating models to study various system behavior is a long/well-known tradition in science. Models can be miniaturized versions of the real items (like airplane models used in wind tunnel studies to analyze aerodynamic characteristics of the design) or life-size mock ups made from cheaper material to mimic certain characteristics that are of interest in the study. While they are interesting, using one domain model to study a system in another domain is much more interesting from modeling point of view. Of course, SW based simulation models are used everywhere these days. My cousin Ravi designs the core of a finite element analysis modeling SW tool his company sells which is used to generate finite state element models of whatever physical entity you want to build from a cell phone to space ship. While most of us are familiar with all these different models used in everyday life, I always fondly recall a simple astable multivibrator I built using pneumatic circuits on an off-shore oil platform couple of decades back for fun. :-)

If you don't remember or know what a multivibrator is, you can read http://en.wikipedia.org/wiki/Multivibrator. There are cool java based web pages that demonstrate Bistablehttp://www.falstad.com/circuit/e-multivib-bi.html, Monostable http://www.falstad.com/circuit/e-multivib-mono.html and Astable http://www.falstad.com/circuit/e-multivib-a.html multivibrators. Basically bistable multivibrator has two stable states with current flowing on one or the other side of the circuit. If you provide a trigger, it will switch state so that current flows on the opposite side. While Bistable mv (also called flip-flop) will remain in that state until the next trigger is received, monostable mv will return to the original state after some time determined by the resistor and capacitor values used in the circuit. The astable vibrator will keep switching sides on its own periodically without the need for any external stimulus. You can see all these in the web pages listed above, by clicking on the set/reset buttons.

Twentyfive years back when I learned about Bi-stable, Monostable and Astable multivibrators for the first time, I was really excited. I remember designing one new digital experiment to create a digital counter and giving it to the professor. To my engineering classmates' chagrin, he promptly added it to the lab test cycle forcing everyone to do that experiment in that semester's digital electronics lab class. :-)

Using bistable mv to remember 1 (current flowing) or 0 (not flowing) or using astable mv to generate clock pulse are all well known. We used to design simple circuits using the 555 IC that had just one astable multivibrator inside to generate audio freq range (20Hz to 20Khz) pulse and feed it into a speaker which will then make a constant noise. By adding a telegraph switch in between to make/break the circuit, we could use it to practice sending/deciphering Morse code for HAM class. :-) My friends Sundaresh, Natarajan and I used to do about 20 words per minute during the 1985 time frame when we got our HAM licenses. That is pretty good speed. Obviously I have lost that skill completely now due to lack of practice.

Anyway, later when I was working as an Instrumentation maintenance engineer in an off-shore oil platform, I used to deal with pneumatic equipments all the time. Even though we had studied the idea of applying one domain model/knowledge to another domain in school, it started making sense only then. I could understand how air pressure, volume bottle, flow restrictor, tubes, etc in pneumatics are functionally identical to electrical voltage, capacitor, resistor, wires, etc. respectively in electronics. So, on the instrumentation lab table, just for fun, one day I built a pneumatic astable multivibrator http://www.falstad.com/circuit/e-multivib-a.html with my colleague Shyam or Kalyan substituting transistors with three way valves and speaker with an air-horn. In 15 minutes when we finished building it, instead of turning on a switch, we opened the air supply valve in the lab to hear the air-horn making a beep-beep-beep sound. We could easily control the freq of the beeps by adjusting the flow restrictor in that pneumatic circuit. Though obvious, I thought seeing it work was pretty cool. :-)

We all know pneumatic or hydraulic circuits can not open/close anywhere as fast as electrical circuits. But, physically building these models show you how closely one system can mimic another. Later in those years we routinely worked with pneumatic circuit boards (where grooves in the boards form circuits), control panels that function like computers to manage oil wells in unmanned oil platforms using no electricity but just oil & gas derived from same oil wells, etc. Since then I have also read books like "Complexity: The Emerging Science at the Edge of Order and Chaos" describing the kind of inter-disciplinary research work being conducted at The Santa Fe Institute, that included applying models of migratory birds flight to Citibank loan disbursement & recovery..! Currently in my company, we are building an extremely complicated SW simulator to mimic the functionality of an even more complicated multicore communication processor ASIC we are designing. Despite my brushes with such much more complicated models, whenever I see an article like "Water for Money" listed in the beginning, it still throws my mind back to the day we built that simple pneumatic astable multivibrator on that lab table. It was a day of geek epiphany.. :-)

-sundar.