I remember when I was in college, we went to some problem where there was a smart agent who was on the grid of squares And it had to clear the intersections. It was awarded with a score for cleaning. It had been reduced to move forward, it has to be refilled every and every time and in the end, how many squares it is dirty or clean on the grid, its Found the last score on the basis.
I am trying to study that problem because it was very interesting when I saw it in college, but I can not get it online or anywhere on Wikipedia. Is there a specific name for that problem that you know? Or maybe it was something that my teacher came to class.
I am searching for AI cleaning agents and similar things, but I could not find anything. I do not know, I'm thinking that there are some other names in it.
If you know where to find more information about this problem, then I would appreciate it. Thank you. Perhaps a "stigmergy" approach is closely related to your problem.
There is a starting point, and you can find something by searching for "dead ants" and "robots".
Actually: Instead of modeling an exact strategy that you work towards a possible approach, ants (perhaps) collect their dead by depositing them according to a simple rule, such as "If there are dead ants There's a lot of pile, so I bring this corpse here; otherwise, I'll give a new pile ". You can begin with simplifying your 'cleanliness' situation, and see where you go.
In addition, I think that (second?) Appropriate approach can be prepared with a carefully chosen combination of fitness functions, with genetic algorithm:
- 'The number of steps performed by robots
Certainly if the robot' dies' goes out of starvation, then it automatically removes itself from the gene pool Gives, and -Larwin awards:)
You're a very simple genota The party can start modeling that will be 'calculated' in practice. Consider using simple ga by Inman Harvey, then each genes either provide a part of the strategy or a complete behavior. Example: If G is converted to 1, then the robot will try to wander randomly; If Jean B is also converted to 1, then it will give priority to self-charging, unless there is no dirty tile on the distance X or the possibility of float and model. Your mileage can be different but I can assure that it will be fun :)
Comments
Post a Comment