Matlab is crunching the numbers as we speak.

[Tanzoria]

I don't have that power. You do though

[Ater Nox]

lock it.

[Tanzoria]

yes, you can do that

[Louisistan]

Disappointingly enough, after three days of calculating different configurations of the algorithm, the best result was 2483, calculated using 8 sub-populations of 25 individuals, migrating every 50 generations in a 1-D ring structure with a migration rate of 10% selected uniform at random.
The best arithmetic mean was 2869.67, calculated using 8 sub-populations of 25 individuals, migrating every 10 generations in a complete net structure with a migration rate of 50% where the best individuals migrated.
The best median value was 2887, calculated using 8 sub-populations of 25 individuals, migrating every 10 generations in a 1-D neighbourhood structure with a migration rate of 50% where the best individuals migrated.

More optimisation to follow.

[Louisistan]

Ater Nox

[Tanzoria]

how about that then

[Ater Nox]

So are you looking at whether the best result was due to the lower migration rate, the higher number of generations or the random selection next, or trying other methods?

[Tanzoria]

the second one

[Louisistan]

Jan 27, 2015 18:38:57 GMT -5 Ater Nox said:

So are you looking at whether the best result was due to the lower migration rate, the higher number of generations or the random selection next, or trying other methods?

Ain't nobody got time fo dat. I want to hand it in this week
I'm taking notice of that fact and moving on.

In fact, I took those three configurations I just described and applied the next part of the assignment to them:
However, this time a clear favourite emerged:

A configuration with 8 sub-populations of 25 individuals, migrating every 10 generations in a complete net structure with a migration rate of 50% where the best individuals migrate and a tournament selection provided the best solution 2474, which admittedly isn't that much of an improvement. But the mean and median values (2664.34 and 2675 respectively) have improved significantly.

[Tanzoria]

I'm happy for you

[Louisistan]

I'm tempted, yet hesitant to believe you.

[Tanzoria]

I'm always sincere

[Louisistan]

Now you're just lying.

[Tanzoria]

First of all, how dare you?

[Louisistan]

No.

[Tanzoria]

no?

[Ater Nox]

Jan 28, 2015 4:43:22 GMT -5 Louisistan said:

Jan 27, 2015 18:38:57 GMT -5 Ater Nox said:

So are you looking at whether the best result was due to the lower migration rate, the higher number of generations or the random selection next, or trying other methods?

Ain't nobody got time fo dat. I want to hand it in this week
I'm taking notice of that fact and moving on.

In fact, I took those three configurations I just described and applied the next part of the assignment to them:
However, this time a clear favourite emerged:

A configuration with 8 sub-populations of 25 individuals, migrating every 10 generations in a complete net structure with a migration rate of 50% where the best individuals migrate and a tournament selection provided the best solution 2474, which admittedly isn't that much of an improvement. But the mean and median values (2664.34 and 2675 respectively) have improved significantly.

Slacker

[Tanzoria]

why don't you do it then

[Ater Nox]

I have better things to do with my time than Louisistan's assignments.

[Tanzoria]

Like spamming?

[Ater Nox]

Exactly

[Tanzoria]

shame

[Ater Nox]

his loss

[Tanzoria]

Let's discuss Bruce Springsteen

[Ater Nox]

Thoughts?

[Tanzoria]

he's quite good

[Ater Nox]

I'd say he was average.

[Tanzoria]

blasphemy

[Ater Nox]

Not really

[Tanzoria]

yes really