Friday, March 04, 2005

Floating Point Errors and Computational Models

A potentially useful and instructive article on the perils of using floating point numerical representations in multi-agent systems.

3 comments:

Brian H said...

Interesting stuff. It also serves as a reminder that the universe operates in something like an infinite-precision mode (until you hit quantum effects, and even then individual events are operating at infinite precision, STS), and that this may have qualitative and quantitative effects -- one reason I harbour a deep-gut distrust of the Turing Model. :)

Martyn Amos said...

What do you mean exactly by "a deep distrust of the Turing model"? I'm curious...

Peakperformance Ken said...

Very Nice. I am really impressed with the great blog you have here! I was wondering how long you have been doing this?

I have a Nano-Technology
Energy Patch
site/blog. It has some simply amazing testimonials regarding the unique
technology that is good for those from 10-110. Just see if you can believe this Nano-Technology
Energy Patch related stuff. Imagine a 12 year old girl weighing less than 100 lbs establishing the
World Squat record of 275 lbs not once but 3 times !

How about an 83 Yr Old Lady that could not stand and had to be carried back and forth to the bathroom and on a Thursday evening in 15 minutes she was not only able to stand, but also walked across the room just holding onto her daughter's arm. The next night, she and her daughter were arm wrestling and for several minutes were locked in the center position. Her daughter the next day said "That was my mom's WEAK arm" :-)

Come and check out our LifeWave Energy
Patch
site/blog if you are interested in getting more energy and or sleeping better than you have in years, and I think you will be very impressed also :-)