Sunday, April 24, 2011

Immobility in an Over-Constrained System

The most enduring quote in the fiction, Life of Pi, is perhaps: “To choose doubt as a philosophy of life is akin to choosing immobility as a means of transportation.”  There are many things that render us immobile in a philosophical sense.  Perhaps the most prevalent reason why people give up is because they are simply overwhelmed.

There is way too much to try and get right in this day and age.  You need to take care of your kids and/or parents, your significant other, and of course, yourself.  You need to exercise enough, eat right, maintain good hygiene, work, pay taxes, and get vaccinated (not necessarily in that order).  It is also recommended that you find a hobby, spend time with friends, and get involved in your community.  If you are a parent, the to-do list is compounded by things like hockey practice and meet-the-teacher night.  We sign up for so much in life, and then struggle to meet the demands they entail.

Engineering projects can find themselves in a predicament similar to that of overwhelmed people.  When constraints in a project begin to press up against one another, the engineer caught in the middle tends to feel suffocated.  The typical reaction to this is usually the most dangerous one: the engineer is rendered immobile.  I have been caught together with a team of engineers in a scenario of this type, and learned a very valuable lesson from it: indecision is in fact worse than a wrong decision.


I will never forget working on the design of a satellite payload that we will call ‘Immobile 1’.  The payload was unlike any the company had ever produced.  Almost nothing could be streamlined for the design; it was truly a custom job.  The Immobile 1 design needed to be completed in six months, and manpower was extremely tight.  As with most things being launched into space, the design was mass-critical – every gram was being fought over between designers and structural analysts.

About three months into the design phase, it was clear to the team that we would not be ready on time.  The slow motion of the design was caused by many factors, but the principal one was that every time a solution was found for one problem, it posed a new problem.  It was like one of those traffic jam games where ten moves were necessary in order to move the intended car forward one space. 

The project manager was about ready to blow a gasket.  He was constantly fuming over the slow pace of the design evolution, and all of the engineers involved, including myself, felt unbelievably stressed.  Every work day felt like a bad dream.

In linear algebra, we learn that a system of equations can have infinite solutions, one solution, or no solution.  In a typical design, there are infinite possible solutions, but some are more optimal than others.  Immobile 1 was a system with no solution that satisfied all of the equations.  The only way to solve it was to bend the equations a bit.

The design of Immobile 1 started to gain momentum only once the engineers began to commit to certain aspects of it.  Even though these choices were not optimal, it was necessary that they be made.  When you commit to a certain aspect of a drawing, the rest of it begins to become clear.  The next three months were vastly more productive than the previous three. 

In the end, the project was a bit late, a bit heavy, and a bit expensive, but it was completed. 

You can only question a problem for so long before beginning.  Often, moving a complex design forward begins with a commitment to some part of it.  Once things are moving along, new challenges are somehow easier to tackle.  It is like the coefficient of friction, which is lower for bodies that are already in motion.

On Immobile 1, we surveyed the problem for too long.  We tried to optimize too many variables too early on in the design.  Optimization can only take place once a baseline has been established.  Creating a baseline design often involves making consolations, particularly when the project has so many constraints.  Our team was relieved to have completed the design, and today, Immobile 1 is very much mobile, orbiting the Earth several times per day.

Maybe the lessons learned in this over-constrained project can be applied in our over-constrained lives.  At some point, we need to realize that life will not be perfect, or even optimal.  Life can be messy.  We cannot solve all of our problems at once.  If your entire house is a mess, it will not be cleaned in one fell swoop.  You need to commit to one room.  Do not doubt whether the whole house can be cleaned, because as Yann Martel warns, this may prevent you from even beginning to clean it.

When we feel overwhelmed by life, the worst response we can have is to shut down.  Life is a system of equations with infinite solutions.  Arriving at your particular solution will take you a lifetime, so there is no need to rush.  Solve it at your own pace, but do keep moving.   

No comments: