Is evolution a greedy search algorithm?

[Mr. Ignorant]

I'm not sure if the title covers the content, but I couldn't think of any better title without it being too long.

Many people say everything is possible with evolution given enough time, but isn't evolution alone really limited in some way? If so then is it really so self-explanatory that humans as complex as they are, could have evolved from single-celled creatures? Isn't it a little too coincidental?

Evolution always favors the organisms with the highest fitness at a given time, therefore it is a greedy search algorithm (http://en.wikipedia.org/wiki/Greedy_algorithm).

For those who know about search algorithms and in particular this one, my point will be obvious. Greedy search is very limited in its capabilities to find an optimal solution. It doesn't care about the context. It just keeps doing what seems profitable locally, but globally this is usually not profitable. If greedy search fails in so many computer applications, why should we expect it to work in reality? In biology? Is it really likely that humans have evolved by coincidence? If it's unlikely, doesn't it make you think about alternatives? Not necessarily a god. It could somehow be directed evolution because of some law in nature that we do not know about yet.

P.S. I'm not a creationist or anything like that. I'm agnostic.
P.P.S. I may be wrong. This is just something that occurred to me. I have some experience with search algorithms, but I may be entirely wrong. If so please correct me.

[Graybeard]

I'm not sure if the title covers the content, but I couldn't think of any better title without it being too long.

I think it is a good title .... and correct

Many people say everything is possible with evolution given enough time, but isn't evolution alone really limited in some way? If so then is it really so self-explanatory that humans as complex as they are, could have evolved from single-celled creatures? Isn't it a little too coincidental?

I find it self explanatory ... If single celled creatures could have evolved ..... then complex organisms are a given.

I think you should peruse this post CLICK

Greedy search is very limited in its capabilities to find an optimal solution. It doesn't care about the context. It just keeps doing what seems profitable locally, but globally this is usually not profitable. If greedy search fails in so many computer applications, why should we expect it to work in reality?

I'm not sure what you mean by 'greedy search fails in so many computer applications' ..... wouldn't the definition of 'fail' be subject to what you consider 'succeed' ?

If there is no pre-motive or direction then how can something fail ... it can only fail if there is a premeditated expectation. As well, tho you say 'greedy-search' is limited, you give no examples of this limitation .. If something is profitable locally and not profitable globally, then it will be selected against in global use ... and only those selections that profit globally will continue ... all else goes extinct if it selected against.

Natural Selection makes no claims to be successful in outcome. If it is it continues, and if it isn't then it adapts. It is able to adapt because of, and not in spite of, the fact that it is never a perfect fit. Error is a necessary part of successful selections.

cool bananas ... greg

[SteveA]

I'm not sure if the title covers the content, but I couldn't think of any better title without it being too long.

Many people say everything is possible with evolution given enough time, but isn't evolution alone really limited in some way? If so then is it really so self-explanatory that humans as complex as they are, could have evolved from single-celled creatures? Isn't it a little too coincidental?

Evolution always favors the organisms with the highest fitness at a given time, therefore it is a greedy search algorithm (http://en.wikipedia.org/wiki/Greedy_algorithm).

For those who know about search algorithms and in particular this one, my point will be obvious. Greedy search is very limited in its capabilities to find an optimal solution. It doesn't care about the context. It just keeps doing what seems profitable locally, but globally this is usually not profitable. If greedy search fails in so many computer applications, why should we expect it to work in reality? In biology? Is it really likely that humans have evolved by coincidence? If it's unlikely, doesn't it make you think about alternatives? Not necessarily a god. It could somehow be directed evolution because of some law in nature that we do not know about yet.

P.S. I'm not a creationist or anything like that. I'm agnostic.
P.P.S. I may be wrong. This is just something that occurred to me. I have some experience with search algorithms, but I may be entirely wrong. If so please correct me.

A greedy search algorithm could be a good analogy for evolutionary theory.

You can even generalize upon that to include any form of deterministic algorithm ... they're all effectively identical in describing ... no, not linear evolution, but instead cyclic systems.

Take any set of rules for state changes within a system that has a finite set of possible states. Describe any rules for change from any state to any state (and yes, you could even include unreachable states in that) and at some point a state transition will have to return to a previous state - again, that's for a finite/enumerable set of states and for deterministic transitions, which would require that all the information regarding a change in state be present in a current state itself.

.... now forget all that because that appears to fail to describe how reality could actually operate and I can point out some paradoxes that seem irresolvable. The main one is that there would be no manner to actually witness anything repeat because there would be nothing within that finite system to distinguish between multiple passes through an identical state. If all information was describable by the state alone then, for example, state #1 could not later because state #1 (that's repeated the 2nd time).

There are some other issues as well - for example, could any form of memory of past states exist within the system itself? Also, there's no way to place information regarding a "present/here/now" state within that.

Anyway, I spent some years trying to resolve some of those issues but never managed to get things to piece together without problems and the resolution for me has been to accept that there's something infinite around and that opens up a ton of interesting possible phenomenon that actually seem to correlate well with experiences, though it also tends to imply that macroscopic physical laws are malleable to an extent and don't necessarily have a figurative "inherent inertia". One the other hand, I'd still assume that for all intents and purposes quantum mechanics describes the general character of how multiple possible growing evolutionary paths could be superimposed in a space, though it seems a question whether or not conventional evolution on Earth is really fundamental or if there could be layers of evolution "below" or more fundamental than that observed, or even if multiple evolutionary pathways could occur in parallel (similar to parallel worlds ideas in quantum mechanics) and evolutionary pathways

Also, you mentioned evolution could be described as a greedy algorithm relative to some fitness function, but then the question could become one of whether or not that fitness function itself was static or supportive of a greedy algorithm. It could be that the fitness function is something incompatible with a linear evolutionary strategy, in which case trying to interpret it as a conventional "function" would be misleading and I think that's related to your comment that greedy algorithms give static/finite/local solutions.

Something else to consider would be to what extent evolution could be tied to individual species as well as life as a whole. There might be aspects of evolution in which survival of one species requires the presence of other/many species and then evolution could exhibit properties of a coordination between many forms of life (there seem to be many examples of this in nature where some actions could be considered innate/instinctive/subconscious/collective. Do those arise from selective evolutionary pressures over time, or is it a distributed property of life itself, or could either view work equally well and be interchangeable or considered simultaneously true and neither wrong?).