Modeling life.

DISCLAIMER: The following article is intended for nerds. Any sane person reading this will probably experience headache and/or depression after reading it.
Prerequisites: Thermodynamics 101 and 201, Optimization 101, Knowledge/Experience in real life should be minimal, Loser preferred.


To model life as a complex energy landscape in a multidimensional space is, to say the obvious, geeky. The fact that you are trying to relate academic knowledge to real life situation is esoteric but at the same time, unfortunately, sad. But here I am, trying to put some sense into life.

So, life is a journey through a complex landscape, where the global minimum is a very narrow, very steep, a very dark well called death. Once you get into it, its impossible to get out of it. Things like 'after life', 'rebirth', etc are complex valued solutions to the problem and are beyond the scope of the present article. So like a good optimization procedure, in order to lead a 'good' life, one shouldn't get stuck in a local minimum for too long. You need to overcome those barriers in order to sample over a larger range of the multidimensional space. Then again, life is not about the destination (the inevitable global minimum), but about the journey. So its also a calculus of variations problem, wherein, you want to find the best path to your death. What happens if you get stuck in a local minimum, one might be tempted to ask? The answer is obvious, you will tunnel through to the global minimum, with almost all of your path spent in an imaginary complex space. And that may not be the 'best' way to die. :)

The success of stochastic optimization methods like Simulated Annealing/Monte Carlo over the conventional gradient based methods (like steepest descent, conjugate gradient, etc) has an interesting implication on the problem at hand. They suggest that a planned and methodical approach may not be the best way to lead life. The problem being, it is too easy to get stuck in a local minimum and subsequently, lack the energy to overcome the barriers. We need to shake things up a bit or take some random decisions which will help us go uphill and hence overcome the energy barriers to allow us to sample a larger space. The role played by 'Temperature' in these methods can be very easily extended to the present problem.

Now, if you spend most of your time away from the local minimas, then equilibrium thermodynamics may not be the best way to go about things. It would be wise to use non-equilibrium thermodynamics. Assuming linear regime of operation, the fluxes are directly proportional to the driving forces. Fluxes being related to your motion in the multidimensional space. The driving forces are pretty obvious. But the irony of the matter is that these driving forces all try to take you towards the local minimum, which you so desperately want to avoid. Herein lies the efficacy of your optimization methods, you need to overcome the driving forces which pull you towards the local minimum.

All the above stuff puts a lot of burden on the person concerned to find the optimal solution. Which is not a good thing for slackers/trivializers like myself. But, once again, science comes to our rescue. Basic postulate of irreversible thermodynamics states that the evolution of a system takes place such that the rate of entropy generation is always positive and the path followed is one of maximum rate of entropy generation. Which basically means that, no matter what you do, the chaos and disorder in your life will always increase. So, chill and enjoy life. :)