hello, I would like to create a simulation about the ants ''s behaviour. I would like to know if someone has knowledges about that or know some web sites where I could find information. thank you for your help!

I would like to create a simulation about the ants ''s behaviour.

that sort of thing will make you sooo much stronger of a programmer then following recipes/tuts/biting off huge game chunks. find things that you are interested in.

you want to bring ant behavior to a simulation. think your buzz word that you might find interesting is "ALife". if you knew that sorry.

google.com had intersting results but i cant say i know the turf.
http://www.google.com/search?hl=en&q=Alife+Ants

Long post ahead, but some good stuff (I think)...

I''m sure you could find some good biology/zoology books on ant and/or insect behaviour. The main emphasis on ant behaviour is social interaction and group based behaviour, a single ant on its own is not too intelligent, but the colony as a whole (like a living organism) is!

----- ANT PATHFINDING --------

For pathfinding, Ants lay down pheremone trails (as more Ants move along a certain path, the stronger the pheremone trail). Pathfinding around an object is done purely by trial and error, but once the pheremone track is strong enough it will be maintained. If the food source at the end of the path dries up, then so does the pheremone trail, and a new trail will be laid down to a new food source. Ants which do not adhere to the pheremone trail much, and vere off can be regarded as less intelligent (and maybe die off, allowing a genetic evolution of your colony to become stronger!)

------ ANT COMMUNICATION ----------

Leaf-cutting Ants (found mainly in South and Central America), also communicate by rubbing a file against their abdomens, to create a sound (audible to humans). They also carry out a form of division of labour to tend their mushroom gardens. Some ants find leaves, some maintain a compost heap, others expand the nest.

Further research shows a wide range of signalling abilities. Including signals for:

* warning
* attraction
* grooming
* recognition
* discrimination
* caste determination
* territorial indicators
* sexual communication

The medium for all of these signals is also chemical based, ''it appears that a colony operates most effectively when its members are genetically tightly knit''. Therefore, an Ant which strays onto foreign territory is killed by the other colony (for those gruesome ones out there, it is held by its mandibles whilst it is stung or sprayed with a toxic substance such as formic acid!)

------- OTHER FACTS -----------

studies also indicate that orientaion is also achieved through memorising visual landmarks, they also make use of the suns location for navigation when the pheremone trail has been removed by sandstorms or winds.

Most of this stuff has been taken from a book written by my tutor at uni:

"QI: The quest for intelligence"
By Professor Kevin Warwick, Department of Cybernetics, University of Reading

That should be plenty to get you started, along with a little imagination, you should be able to come up with a good start.

I''ve developed a more crude system of cells which genetically evolve, based on three types of cell, varying in strength, and with mutations being generated which actually kill weaker cells, which not only strengthens the overall conlony, but also prevents the colony growing at an exponential rate, and keeps it at some reasonable optimum. You might want to consider either predators, or severe punishments for un-intelligent ants.

Hope this helped, and if you''re still reading, then sorry it was so long!

cheers

Matt

For a interesting simulation you would want to make each ant a seperate entity, I think the word is single celled automaton. So don''t have a outside class that just tells all the ants to walk in a line towards the food, but give each ant a few possible states one of which is current, and a few actions and communications. After programming what a individual ant should do, then throw a bunch of them together in a simulator with food, tunnels and a queen ant and try to see if they can work together. Then tweak the ant class code until they will work together nicely. Really when you program what goes through a ants mind and what influences his decisions, you probably don''t need much technical info, you just have to know what a ant priorities are, you can almost bet they want sex first if horny, then food next with foods with higher suger being preferred, and then have to rest every so often. And if the chunk of food is more then 500 times the ants weight, it must break it into smaller pieces to carry. You also want to make sure they can tell other ants a direction of a food source or say follow me.

ya the general thing to keep in mind is its the little rules making complicated behavior.

dont have a hive class that decides that we need to expand.

tweak the rules for the individual ants then see what happens in mass.

hi,

thank you for your answers.


----- ANT PATHFINDING --------

For pathfinding, Ants lay down pheremone trails (as more Ants move along a certain path, the stronger the pheremone trail). Pathfinding around an object is done purely by trial and error, but once the pheremone track is strong enough it will be maintained. If the food source at the end of the path dries up, then so does the pheremone trail, and a new trail will be laid down to a new food source. Ants which do not adhere to the pheremone trail much, and vere off can be regarded as less intelligent (and maybe die off, allowing a genetic evolution of your colony to become stronger!)

this is exactly what I would like to simulate. after a while, we should be able to see what is the shortest way between the colony and the food. I m going to read several articles about taht and I will give a feedback.

thank you

ohhhhh crazy ant behavior story...

ants use the shortest time path not the shortest physical path.

i was walking along one day heading to school bout two years ago. i notice this string of ants. now this string was running along the edge of some grass on the concrete right. but the concrete took a corner. a walk way up to a church.

but picture the scene the ants werent using straight line shortest distance through the grass. if they had been it would have been atleast a 45 degree junt off into the grass to clip the corner since the path was strong 10 feet in each direction along the concrete.

conclusion ants use fastest routes not shortest distance. but how does an ant figure out which way is faster? im sure its not magic. but its still interesting.

Ants are already used as metaphore for optmization tasks in research, see Ant Colony Optimization.