GA results with gnuplot

Tests are done on a MAC book pro (2.3G dual core) and iMac (3G dual core) 


With small population sizes and target reached at 98% execution time < 5 seconds.


Notice here that only with small populations we have several fallbacks on strength and a very unstable curve due to greater impact of stochastic mutations and crossovers on the intermediate average strength:





With medium population sizes and target reached at 98%:(execution time < 1 minute)


Here you can see that evolution stands still for a number of generations and only by a random mutation in time evolution starts climbing again:

With large population sizes and target reached at 98%:(execution time < 2 hour)

Remarkable here is that in the first 100 generations evolution is going to 90% of the target with no fallback at all!

Psize = 10      => [ 2 .. 10 ] seconds to execute 
                => order of generations needed [ 10000 - 80000 ]
Psize = 400     => [ 30 .. 60 ] seconds to execute 
                => order of generations needed [ 300 - 1000 ] 
Psize = 1000    => [ 120 .. 300]  seconds to execute 
                => order of generations needed [ 100 - 300 ] 
Psize = 100000  => [ 6000 .. 10000 ] seconds te execute

                => order of generations needed [ 100 - 300 ] 
Below you can see this exponential increase of the execution time vs the population size:





Also the generations needed decrease exponentially with an increase of the population size: