network profile
| bio | website | nextlevelsolns.com |
|---|---|---|
| location | ||
| age | ||
| visits | member for | 1 year, 3 months |
| seen | Mar 8 '12 at 11:31 | |
| stats | profile views | 0 |
|
Feb 12 |
comment |
Convergence of a genetic algorithm BTW, does anyone have a reference for this approach? |
|
Feb 12 |
comment |
Convergence of a genetic algorithm Thanks, this is what I'm doing now, in fact a little more. If I don't see any improvement in 10 generation, I keep the single best model, toss out and rerandomize all the others. I thought this was working well, until I just had a model that found a better solution after 45 generations, and the 2nd rerandomization. I'm working on implementing the multiple chains approach used in MCMC. Start 2 or 3 chains in parallel, when all 3 chains agree on the best 3 or 4 models, it is done. |
|
Feb 11 |
awarded | Student |
|
Feb 10 |
asked | Convergence of a genetic algorithm |
