"The Greedy Planet", nice title that, grabs attention. Is this going to be a rant bemoaning how the world has come to become a greedy place? No, on the contrary, this post considers the idea the world is inherently a greedy place. I am not talking about morality here, instead I am talking about the world "running" a "greedy algorithm". Sounds geekish? Bear with me for a couple of passages and I will elucidate what I mean.

Water flows down, more generally, fluids tend to flow downwards to lower their potential energies under influence of gravity. If you don't mind me anthropomorphizing water, let me ask a rhetorical question: What is the "purpose" of water? Does it seek to find the bottom-most point on the planet so as to reduce potential energy to a bare minimum? The answer seems to be "yes", for, given an empty river bed from wherever the water is to wherever the bottommost point on the surface is, water would naturally flow to the lowest point.

>Why does water need something like a "river bed" to fulfill its "purpose"? We can answer this question if we explore the "algorithm" water uses to realize its "destiny". Water follows a simple "algorithm": from the current position go to a neighboring point that has the least altitude, repeat this step until there is no neighboring point lower than the current point. This kind of procedure to solve a problem is generally called a "greedy algorithm".

A "greedy algorithm" is not an "optimal algorithm". This explains why all the water in the world doesn't somehow find its way to marina's trench and "realize" its "stated destiny". Flow of water suffers from a classic problem of a "greedy algorithm": getting stuck in a "local minima" (or "local maxima", if the algorithm is trying to maximise something).

Let me make clear the concept of "local minima" (skip this paragraph if you know it already). Lets say you are standing on top of a mountain that features a precipice and an empty lake midway down to the bottom. You pour a some amount of water from the summit. The water flows down the precipice and comes near the lake, which it enters and stays put. Had the water "decided" against "staying" in the lake and crossed over the edge of the lake, it would have done a lot better in terms of "losing the potential energy" target, because it had halfway more of the mountain to climb down. The water in effect got stuck in the "local minima" of the lake instead of exploring further to find the "global minima" at the base of the mountain (or further down).

"Greedy algorithm" thus is inherent in flow of fluids.

Lets see another example, the "algorithm" of evolution of life: "Survival of the fittest". This puts a simple criterion for survival of species: if you can survive your immediate environment: you live, else you die. This too is essentially "greedy". Lets say you could evolve a truly magnificent species [a species that can survive a wide range of challenges] through your current generation of species, but the evolution has to go through a "weak link", I.e., intermediate species that can't survive on their own. A truly "optimal" algorithm of evolution would follow this path "somehow". But for "survival of the fittest" the "weak link" is the dead-end. In essence the species before the "weak link" is the "local minima" for the "survival of the fittest" "algorithm".

Are we humans "naturally" greedy? Well, I think answer now enters the fuzzy grey areas between yes and no. We are surely not as "greedy" as "survival of the fittest" or "flow of water". But we surely aren't anywhere close to being "optimal".

Economic structures give us a good example of how we structure some aspects of our lives. The sole "tried-and-tested" economic system that we know works is "capitalism". This really is a watered down version of "survival of the fittest". The system revolves around "competition" between firms and individuals in what can essentially be approximated as a "free" market. If you can't stand up to the competition, you "die" (economically), if you beat your competition, you "survive" and "pass on your genes". Of course, a lot of "unfairness" of "survival of the fittest" becomes "illegal" in "capitalism". This is the reason why I called it a "watered down version".

There is an interesting example of an attempt at building a "non-greedy" and "optimal" economy: The USSR. The USSR, in principle, sought to do what is good for the proletariat. They could make tanks instead of making shoes if that is what the country "really needed", even if the immediate market demand was for shoes: a non-greedy algorithm for sure. The attempt failed miserably. I think this is because of three reasons:
1) There isn't a substitute for "individual greed" as an "incentive" to make people work harder. The idea of "common good" failed pretty badly here.
2) The "solution space" for planning an economy is humongous. It is possible that the planners couldn't figure out a way to run their economy in the direction of an "optimal solution".
3) The "individual greed" of the planners was never suppressed, this messed the system that was supposed to run on "common good".

If you scrutinized the way I headed in this article, it must have been pretty clear to you that I've argued that "greed" is "inherent" in "algorithms" that "run" a lot of natural systems. It is even inherent in most of the systems which we use to organize ourselves. Of course, the "algorithms" of "life" aren't "truly" greedy, they are more "co-operatively greedy". Game theory [1] in general and prisoner's dilemma [2] in particular can shed some light as to why "co-operation" is deemed "desirable". The "co-operative behavior" notwithstanding, it must be noted that most of our systems really "run" a "greedy"-based heuristic and we are nowhere close to "running" a truly "optimal" algorithm.

[1] I'll probably come back to this in one of the future articles.

[2] You might also want to read the book with the same title. Link to Prisoner's Dilemma on Amazon.

Created on 1^st August 2006

• Home
• About
• Links
• Credits
• GuestBook
• E-Mail Me

• Technology
• Book Reviews
• C In DOS
• Reason

• Projects
• Cool Ones
• Headers
• TSRs
• Others
• Syllabus Programs
• User Files
• VTU Projects

• Standard
• Boxy
• Autumn
• Original