Another low variance long running session

This will probably be the last parameter optimizing session, as parameter optimizing does not bring the expected results.

The average fitness never exceeds 1000. After 1000 generations there is an improvement in the maximum values, but the no further advancements.
Have a look at the robots after 0, 300, 1000 and 2216 generations.

A long running low variance breeding session

This breeding session generated the absolute maximum average (and max) fitness. Relatively fast it rises to more than 10000 in average. That means that more than half of the individuals of a population could win their matches against the reference robot.
Why could there no further improvement? No idea.
See the robots in action after:
0, 1000, 2000 and 2384 generations

Low variance initial generation

As the performace of the results of parameter optimizing did never get even near the results of the manually designed fuzzy controllers, I tryed to make the inital generation almost equal to the manually choosen parameters (low variance). The GA sould not find a better solutioion somewhere in the waste area of parameters, but find an optimized solution staring at a almost optimal solution.
Here are the results of that.

The fitness rises immediately to a much higher average and maximum than in the sessions with random initial generation (see the results in older posts).
Here are some of the best performers of their generations. The violet sumo is the 'breeded' one.

• 0 generation
• 1 generation
• 100 generations
• 500 generations
• 800 generations

Another long running breeding session

See the behaviour after several generations.

• 1 generation
• 370 generations
• 1354 generations
• 2168 generations
• 3066 generations
• 4116 generations

A long running breeding session

I let a session run for several days (on my not very fast computer) to see what happens after several thousand generations.

The result of that session shows that a minimum improvement can be achieved after a view hundred generations.
Check the behaviour of the fittest individual in that phase.

• 0 Generations
• 188 Generations
• 1943 Generations 

The pink sumo robot contains the 'breeded' controller whereas the blue one is controlled by a manually designed fuzzy controller. Obviously the breeded robot does not have a chance towin against the manually designed one. Even not after more than 1900 generations.
Here are some more breeded robots after the increase of fitness at about 3500 generations.

• 3863 Generations
• 4089 Generations

Another Parameter Optimizing Result

This was a quite long running breeding session. You can see that relatively early (about 200 generations) the maximum fitness exeeded 10000. That means there where individuals could defeat the eference controller. This is a good result, but from that time on no further progress could be measured.
I do interprete that result in the way that winning a game must have been by chance. Further investigation in that direction will follow.

The beginning of that session.

Sumosim in Action
To see what is the behaviour of the best breeded robots click on one of the following links.
The breeded robot is the violet one.
- Generation 310 webstart...
- Generation 1741 webstart...

First Results of the Parameter Optimizing GA

What does that mean? You see the fitness of a whole population of 50 individuals for diffrerent generations (about 180). The red line is the average fitness of all individuals, blue shows the worst - and green the best fitness.

How is fitness defined? Each individual has to play ten times against a rather good fuzzy controlled robot. You can see two of these robots in action here.... If the tested controller looses it gets 900 points, if the game is undescided it gets 1000 points and if it wins it gets 1100 points. The results of the ten games gets added. Aditionally to the points there is a bonus for the duration of the game. If the robot looses, but the game lasts long it gets a relative high bonus. If the robot is the winner it gets a bonus for short duration of the game. So the fittest is if it wins in short time.
A fitness of 9000 means that the tested robot looses always in very short time. After about 100 generations the max fitness gets values that are sometimes 9800. That means there was one robot that could avoid to be kicked off the arena for quite a long time.
Hopefully the max value will exceed 10000. That would mean there were robots that could defeat the testrobot.

Sumosim in action

To see how sumosim works visit the following link: http://entelijan.net/sumo-sim.
Click one of the numbers below examples to start a simulation inside a java applet