Advertisement

how do players make decision in a football/basketball game?

Started by June 01, 2006 04:07 AM
0 comments, last by Timkin 18 years, 5 months ago
I have browsered a lot of files about sports game. I think players's state can be controled with a FSM.like this state ,message ,next_stage , execute_function state_wait ,msg_ball_come ,state_keep_ball, catch_ball_animation() state_keep_ball, msg_please_through_ball,state_run, throw_ball_animation() ........ How to generate a messge like msg_please_through_ball?a.k.a :how do players make decision? for a state of running ,how to decide a target?I think there maybe have a dynamic balance map. anybody can help me or give me some advice or doc? thank u
Quote: Original post by cantonaNJ
state ,message ,next_stage , execute_function
state_wait ,msg_ball_come ,state_keep_ball, catch_ball_animation()
state_keep_ball, msg_please_throw_ball,state_run, throw_ball_animation()
........

How to generate a messge like msg_please_throw_ball?


Some messages would come from the environment while others would come from other players. Environment messages tell the agent information about obstacles (like other players, walls, goals, etc.,). Messages from other (AI) players are related to gameplay. So this indicates two levels in your AI: (1) the ability to move around the game field to destination locations while avoiding obstacles; and, (2) the ability to interact with other agents to achieve a gameplay goal.

Your specifc question seems to indicate how AI players generate the latter. There are two principle paradigms for cooperative agents: (1) all agents are guided by an oracle (a single AI that makes decisions for all); or, (2) all agents choose their actions based on how they perceive the world (and possibly after negotiating actions/goals with other players).

As for the algorithms you use to achieve this, it depends on which paradigm you choose. (I'd suggest you start with Decision Trees)

Quote: for a state of running ,how to decide a target?I think there maybe have a dynamic balance map.


This brings up the issue of strategy versus tactics in team play. It's one thing to have your AI coordinate player movement and ball passing to set up a shot on goal; it's another thing entirely to do it while the opposing team is trying to take the ball from you and simultaneously crowd their backline. You need to be able to analyse the game field and extract relevant strategic information. A good starting place for this kind of analysis is the Influence Map. There's plenty of literature available, although you might like to start by reading some of Penny Sweetser's work (University of Queensland).

Cheers,

Timkin

This topic is closed to new replies.

Advertisement