sado / kewe: (parameter) SPACE EXPLORATION

[Lumphie]

I ran 100 simulations of the sado model (used swap, lut and no bug to make it quicker)

All parameters were the same as the sado article except for sigma c, which started at 0.01 (first simulation) and ended at 1.0 (100th simulation)

I created this wonderful GIF of the ecological type.

Female preference is still uploading.

Will make more simulations that increase other parameters and see what works.

[Lumphie]

Ecotype branching as in the article happens somewhere around 0.6-0.7

For the next exploration of parameters space simulations sc will be set to 0.6

[Lumphie]

Female preference with increasing sigma c. Not really interesting.

[Lumphie]

runs with sigma e are just in. Not really interesting.

Best around 0.5-0.6 probably.

For the next exploration of parameters space simulations se will be set to 0.5

[Lumphie]

On peregrine now: sk from 0.02 - 2.0 (Simulations12) sm from 0.005 - 0.5 (simulations_sm)

[Lumphie]

The queue is taking too long. Will check back tomorrow!

[Lumphie]

Higher sigma k => more and more stable branching for ecotype. Maybe more branching on sexual traits, but not a very high effect.

gif is incoming

[Lumphie]

As expected. Biggerresource distribution gives more room for branching.

Will stay with 1.2 sk for now with the knowledge that I can increase it whenever needed.

No (or small) effect on sexual traits

[Lumphie]

Something went wrong with the sigma_m simulations. Will rerun them after sigma_q is done.

[Lumphie]

No fancy gifs of sigma Q.

Nothing interesting. We'll keep it at 1

[Lumphie]

Nothing interesting in sigma_m I keep it at 0.2

Will run multiple simulations of

sm 0.2 sq 1.0 sk 1.2 se 0.5 sc 0.6

Overlapping should have branching. Separate generations does not according to the gui.

[Lumphie]

Okay lets just use the same exploration techniques for separate generations

[Lumphie]

So many problems... euhm challenges.

Forgot to scale c to the population size. It was stalled at 1700 individuals. It should run now with scaled c

[Lumphie]

Note to all: All computers reset at 22:30.

Something I just found out.

[Lumphie]

Increasing C (not sigma c) increases eco type branching with seperate generations. Now I just need a way to increase the sexual traits branching

[Lumphie]

Okay another one, because they are so pretty.

Increasing c (the constant) increases the chance of branching, but it also increases the chance of nobody surviving.

[Lumphie]

And this is female preference.

Right now increasing population size is running and sigma m

[Lumphie]

increasing sm to 1.0 gets very close to branching on female preference, but somehow not on male trait...

Which is of course because making sm higher lowers selection pressure on the outsiders of female preference, but not on the outsiders of male trait.

[Lumphie]

Female preference for an increasing sm.

[Lumphie]

300 simulations running right now!

different population sizes sq and 100 random simulations with the following parameters

sc 0.6 sk 2.0 sv 0.02 se 3.0 (!!!) pop 10,000 (!) c 0.000005 sm 1.0 (!) sq 1.0 (I expect this should be enough)

Only time will tell what'll happen.

[Lumphie]

Okay so the 100 random simulations are in. Not looking like it's enough.

In order:

Eco_trait Fem_pref Male_trait

[Lumphie]

Can't get the male trait to follow the females. sq might be too high.

Eco type is not getting big enough, so higher Sk and probably a higher Sc (to get the branching hopefully)

[Lumphie]

Or maybe increase c a bit.

[Lumphie]

Now running:

sc 1.2 sk 4.0 sv 0.02 se 3.0 (!!!) pop 10,000 (!) c 0.00005 sm 1.0 (!) sq 0.5

[Lumphie]

Also

sc 0.6 sv 0.02 se 0.5 pop0 1000 c 0.0005 sm 1.0 sq 0.6 type0 0 0 0' > kewe_sado_parameters${i}.txt

echo "seed ${i}" >> kewe_sado_parameters${i}.txt

sk=$(printf '%s * %s\n' "$i" "0.05" | bc) echo "sk $sk" >> kewe_sado_parameters${i}.txt

Increasing sk (0.05 -> 5.0 with steps of 0.05)

[Lumphie]

If eco_trait branches anyway a low se should let the sexual traits follow.

[Lumphie]

So I have to get ecotype to branch first.

Starting at sk this time.

sm 0.2 sq 1.0 sk 1.2 se 0.5 sc 0.6

These gave branching at both ecotype and sexual trait in the overlapping generations. Running increase sk now (starting at 0.05, ending at 5.0)

[Lumphie]

increasing c looks promising:

ecotype fempref maletrait

[Lumphie]

ONE DAY WE WILL FIND THOSE PARAMETERS! BUT IT IS NOT THIS DAY! THIS DAY WE USE PEREGRINE. FOR SADOOOOOOOOOO!!!!

[Lumphie]

Running simulation for 40.000 generations for around 10.000 individuals and a resource distribution of 10.0.

Lets see what happens.

[Lumphie]

Simulation got stuck because population size kept increasing. Increased c and ran it again.

It's not at generation 20.000 so halfway! :smile: 14.000 individuals now.

[Lumphie]

Damnit

[Lumphie]

I think I found some parameters which have speciation. Have to get the histogramwidths right though.

[richelbilderbeek]

It has been done.