# It’s Amazing What Evolution Can Do!

October 25, 2010 | Posted by PaV under Evolution, Intelligent Design |

This article here recounts the now documented ability of bees to solve the “traveling salesman problem” faster than computers. And to imagine that evolution has done this! My, what a wonderful thing it is!—-(he says with sarcasm dripping). By just doing something over and over again, with little changes accumulating, a ‘computer,’ better than any we have, somehow comes into existence. And, of course, this ‘computer’ is the size of a grass seed (!!). One of the experimentalists said this: “Despite their tiny brains bees are capable of extraordinary feats of behaviour. . . We need to understand how they can solve the travelling salesman problem without a computer.” I agree with his statement. I would only suggest that RM+NS won’t lead to this understanding.

### 21 Responses to *It’s Amazing What Evolution Can Do!*

### Leave a Reply

You must be logged in to post a comment.

There is nothing ‘evolved’ about the honey bee:

Evolution Vs. The Honey Bee – an Architectural Marvel – video

http://www.metacafe.com/watch/4181791/

50 million year old honeybee:

http://www.fossil-museum.com/f.....hp?Id=9643

http://www.fossil-museum.com/f.....asonuc.php

No, the bees did not solve the traveling salesman problem.

To solve the TSP is to provide a mathematical solution to a formal problem. The bees, however, are demonstrating a pragmatic solution to a physical problem. That’s quite different.

I’ll illustrate the point. Perhaps the bees wanted to visit 20 flowers, but couldn’t solve the TSP. So they simplified their problem and visited only 10. Then the investigator examines what they actually did, but is not aware that they pruned to problem down to a simpler one. So he describes only the simpler problem.

That a problem is described as the TSP doesn’t show that it actually was the TSP.

This is presented as a story on “how smart bees are.” I don’t doubt that bees can do some clever stuff, but the story is better seen as “how gullible science reporters are.”

of related interest:

New Caledonian Crows Exceed Apes/Chimps at Trap-tube Experiment – video

http://www.metacafe.com/watch/4181775/

As usual, it will be necessary to wait for the real article (Lihoreau M, Chittka L & Raine NE. 2010.

Travel optimization by foraging bumblebees through re-adjustments of traplines after discovery of new flower patches.American Naturalist: in press.)BTW: it seems that the travel optimization is used only on the location of the

flower patches– and how many of these are there an a single flight? Not that many, less than ten probably. Perhaps the bees use a kind ofant colony optimizationTo both Neil and DiEb:

If the problem was so simple, then why do the scientists say that it would take a computer days to figure out the right optimization?

Computer simulations most generally involve iterative techniques, and so even with a small number of unknowns to be solved for, the equations can be quite difficult when solving them at random–which I would think is how the program would be set up.

You both seem to want to minimize this amazing capability—”it’s no big thing!”—so as, I would presume, to give Darwinism a “Get Out of Jail” card.

Aren’t you just running away from the truth?

PaV,

What is your hypothesis? How can it be tested?

Michael

The original article (and analysis given here) is terrible. Of course bees can’t solve anything that would require a computer days. They can solve very simple instances of the “Traveling Salsemen Problem” that could also be solved by a person in a few seconds.

There are instances of the traveling salsemen problem that would take a computer days or are even impossible to get an exact solution with present computation. These involves thousands of nodes and getting and are way, way harder than what the bees are able to figure out. Though often times even these problems will have approximate solutions that are much easier to find and is what the bees are going to do even for simple problems.

Sorry, typo should have read

“These involve thousands of nodes and are way, way harder than what the bees can figure out”

@PaV

It’s quite unfortunate, but science journalists have a proclivity for hyperbole. Therefore it is better to have a look into the original research.

I doubt that a

singlebee figures out the optimal path on its own, it is the collection of the bees which makes this possible (swarm intelligence). That is why I mentioned ant colony optimization – perhaps we observe a similar process with bees.To PaV:

No, I am not running away from anything. There’s another post on UD today, about mechanism/information metaphors, and I agree that those metaphors are misleading. I disagree with the whole “the brain is a computer” way of thinking.

I can see why you think of my point as opposed to ID, for indeed that kind of thinking does seem to come from taking a design stance. However, this misleading use of mechanistic metaphors is pervasive, even among many who are themselves opposed to ID.

Full disclosure: I do disagree with ID, but that is not what motivated my comments in this thread.

PAV

Isn’t it amazing how the evolutionist science deniers can twist the obvious meaning of an article so that it does not conflict with there religious point of view. Obviously evolution can not explain the impressive capabilities of the a honey bee. The only alternative to evolution is God did it. That is unbearable to the evolutionist so we see these grand contortions of logic to prop up a pseudo science. Thanks for the interesting post, the implications of which are obvious and profound.

Temp Hut:

Then why do scientists bother with computers? And, did you notice how you use an intelligent agent as a means of simplifying an otherwise complex problem?

Neil:

The whole “the brain is a computer” does push things too far, because there is an element of ‘mind’ in the brains of higher animals, and, perhaps, to some extent even in lower ones. Nonetheless, computer scientists are studying how the brain functions so as to build more efficient computers. So, it would seem this is a little bit more than simply an analogy.

Michael Tuite:

Your comment is meant only to insult. Are you so superior to everyone here?

Exactly why am I suppose to have an hypothesis? The question I posed is still a valid question: how, given supposedly ‘random’ variations, did a TINY brain come up with such a capability? Compare a ‘grain of grass seed’ to a rather large supercomputer—and the little seed whips the computer. How do you explain it? What is your hypothesis?

Pav said

“Then why do scientists bother with computers?”

Because certain instances of the traveling salesmen problem are very difficult for even the most powerful super computers. It is also of theoretical interest because the traveling salesmen problem is NP complete. But that doesn’t mean that all instances of the problem are difficult.

For an example, you are in NY and you want to visit chicago end up in LA. You have already solved that problem I’m sure and what the bees can do is much closer to this than the specific instances of the traveling salsemen problem that take computers days.

bornagain77:

In your post at #3: Are you suggesting that it is OK to be a “bird-brain”?

PaV you bet it is,,,

The Trashmen – Surfin’ Bird

http://www.youtube.com/watch?v=uj0GtRI4Ulo

I can’t read the article, but here is the abstract:

So, they are solving the TSP for four nodes: there are six possible ways to visit the nodes, three when you neglect the orientation. What the text is saying is that the bees have an brutal force approach to the problem (

This resulted in the development of a primary route (trapline) and two or three less frequently used secondary routes.), but are generally able to memorize they shortest way found.BTW, in their paper

Trapline foraging by bumble bees: IV. Optimization of route geometry in the absence of competition(Behavioral Ecology, September 29, 2006) they authors Kazuharu Ohashi, James D. Thomson, and Daniel D’Souza used 10 artificial flowers and showed that bees generally don’t follow the optimal path, but seem to be able to memorize earlier traplines.PaV:

By just doing something over and over again, with little changes accumulating,a ‘computer,’ better than any we have, somehow comes into existence.Abstract:

We analyzed bee flight movements in anarray of four artificial flowersmaximizing interfloral distances.This shows again how careful one should be when reading an blog entry on an article in a newspaper which was itself based on the press release on the actual scientific research (which was most probably not written by the scientists themselves, but by the press office of their institute): it is like a play of grapevine where each player is a little bit more sensationalist than the previous one.

Dieb, what a bad mathematician you are. There are 24 different possible routes to visit all the flowers once in a 4 flower array (4! = 4 x 3 x 2 x 1)

Obviously, there are 24 permutations of four nodes. But there are only three closed not oriented paths: if you think of the nodes as a quadrilateral, you can follow the sides (always the shortest path), or a pair of two opposing sides and the diagonals (the longer paths).

Hope that helps.

DiEb and TempHut:

On Wiki, they say that the TSP can be classified as O(n!).

In the experiment, 10 “artificial flowers” were used. 10! (ten factorial) yields 3.6 million different permutations that have to be evaluated. Assuming a computer can evaluate 36 permutations per second, it will take 100,000 seconds of computing time to calculate the path of least distance by brute force. That’s about 30 hours. If the computer can evaluate only 20 permutations per second, then it will need 54 hours to evaluate.

This makes me wonder why you have chosen to select such “simple situations” such as 4 different flower locations? Are you trying to stack the deck? Or are you trying to search for the truth?

How do you explain that bees can solve this problem? They don’t have a computer. They have to deal with lots of locations. How do they do it? Do they do it perfectly? Apparently not. But how can they even approximate it? They obviously don’t use ‘brute force’ methods. What do they do? How do they solve this problem? And, considering that their lives depend on it, it’s rather important that they be able to solve it. What, then, are the implications here for putative random mutations leading to such a capability?

PaV:

I answered at the more recent thread…