I attached a screenshot of the initialization of 2 teams of 2 players each (1&2 vs 3&4), done in a simple way... if you have minimum intuition or practice with the trigger editor, it will be easy. Just remember that "find" is your friend ;)
(I suggest using the "Ally With Shared Vision" setting, mine is "Ally" because i don't want allies share vision in the map i used this trigger)
this vote was hard to pick, the other three works are surely original and interesting, but the one who impressed me most is rrowland's behemoth, very challenging. I think that once it will be improved, it should be able to beat even koreans :P
that's impressive... what did you use to develop all of this? is java based or .net based or something else?
it seems a very though job, how long did it take until now?
I would have joined the competition but I'm studying for exams and I just played around only a little bit with the editor until now... I wrote down some sketches for an AI system that use different strategies, tactics and even micro, but it's just some drawing on the paper i didn't had the time to do something, it would have been funny :D
I saw very good work here, rrowland micro is very nice, and just the idea of mimic ai is original, gj guys! :D
never thought to build something like that... probably because i hate c and pointers :D
however this galaxy thing is a very rudimental and limited one, i expected better things from 10 years of sc2 waiting...
uh, i have a question: are there anything else other than lock to handle shared resources and critical sections? i saw a wait function but it seemed more like the java sleep function...
i don't have time to explain in detail, however a simple way to design AIs is using finite state machines.
The rough idea is this:
first, you check in what state the AI is currently, and for each state you will have different things to do, but the pattern is the same:
you check your "input" (variables, events, changes in the game, etc.) and consequently you decide what is the "output" (what to do with your local or global variables, triggering events or whatever) and what is the next state the AI will swich (including the current state, i.e. we remain in the current state).
Thalassicus wrote a simple example with two states, "advancing" and "retreating", but you could have far more complex scenarios like managing a whole computer player with it, although it would probably become easy to predict computer moves after a few games, like age of empires and other old rts games of the 90s...
Just remember you have to manage the state of the entity you are designing the AI for, and the transitions from one state to another. A good idea is to use pen&paper and draw a graph, with states as nodes and transitions as edges.
Maybe the pic on this wiki page could help understanding: Mealy Machine
0
i think a lot of TD maps fall into the category.
i'm sure these two works, they are my favourites.
Standard Tower Defense
Shrink TD
0
and wow prepaid instead of luxury taxes :D
0
I attached a screenshot of the initialization of 2 teams of 2 players each (1&2 vs 3&4), done in a simple way... if you have minimum intuition or practice with the trigger editor, it will be easy. Just remember that "find" is your friend ;)
(I suggest using the "Ally With Shared Vision" setting, mine is "Ally" because i don't want allies share vision in the map i used this trigger)
0
this vote was hard to pick, the other three works are surely original and interesting, but the one who impressed me most is rrowland's behemoth, very challenging. I think that once it will be improved, it should be able to beat even koreans :P
0
that's impressive... what did you use to develop all of this? is java based or .net based or something else? it seems a very though job, how long did it take until now?
0
I would have joined the competition but I'm studying for exams and I just played around only a little bit with the editor until now... I wrote down some sketches for an AI system that use different strategies, tactics and even micro, but it's just some drawing on the paper i didn't had the time to do something, it would have been funny :D
I saw very good work here, rrowland micro is very nice, and just the idea of mimic ai is original, gj guys! :D
0
@Thalassicus: Go
never thought to build something like that... probably because i hate c and pointers :D
however this galaxy thing is a very rudimental and limited one, i expected better things from 10 years of sc2 waiting...
uh, i have a question: are there anything else other than lock to handle shared resources and critical sections? i saw a wait function but it seemed more like the java sleep function...
0
i don't have time to explain in detail, however a simple way to design AIs is using finite state machines.
The rough idea is this: first, you check in what state the AI is currently, and for each state you will have different things to do, but the pattern is the same: you check your "input" (variables, events, changes in the game, etc.) and consequently you decide what is the "output" (what to do with your local or global variables, triggering events or whatever) and what is the next state the AI will swich (including the current state, i.e. we remain in the current state).
Thalassicus wrote a simple example with two states, "advancing" and "retreating", but you could have far more complex scenarios like managing a whole computer player with it, although it would probably become easy to predict computer moves after a few games, like age of empires and other old rts games of the 90s... Just remember you have to manage the state of the entity you are designing the AI for, and the transitions from one state to another. A good idea is to use pen&paper and draw a graph, with states as nodes and transitions as edges. Maybe the pic on this wiki page could help understanding: Mealy Machine