|
For the fastest supercomputers with a speed of 100 quadrillions
floating point operations per second,
for a 'simplest NP complete' problem such as the satisfiability problem
with an input size of 100, if 100 data instances are taken as one floating
point operation,
the time needed is
T(n = 100) >> 2 ^ 100 (second-level) >> 100000 (year-level)!
So it is unbelievable that some people can solve it within 10 seconds!
With a input size of 10000 or more, it will blow up the world:
T(10000) >> 2 ^ 10000 (second-level) >> 10000 million years.
TSP is the hardest of 'all' of the most often mentioned types, it need some
extra comparison although it may save some repeating works,
the basic rounds of works are not reduced.
Some people say that they can solve it for an input size of more than one million.
You are really great!
Here, T(n) is the time complexity of n, the worst case of n.
Archiver|手机版|科学网 ( 京ICP备07017567号-12 )
GMT+8, 2024-4-27 17:33
Powered by ScienceNet.cn
Copyright © 2007- 中国科学报社