|
Author(s): |
JiaHao Huang, School of Software,Shanghai Jiao Tong University,Shanghai, China
Qian Yu, School of Software,Shanghai Jiao Tong University,Shanghai, China
ShuangJiu Xiao, School of Software,Shanghai Jiao Tong University,Shanghai, China |
|
Abstract: |
Artificial Intelligence is one of the most important parts of video games and the concept of agent is widely used in AI developing. It is important that intelligent agents can be developed efficiently and productively in video games. For this purpose, this paper presents a lightweight and flexible architecture of intelligent agents. The architecture is different from others that focus on intelligence. It removes the extra works such as defining a lot of rules and highly encapsulates each intelligent agent as an individual. Usually, an agent should contain three parts: sensors which percept the information of environment, decision makers which decide the action that should be processed next and actuators which perform actions in the environment. The architecture first uses the concept of game object to refer the intelligent agent in system; second, “components” are added into game object to implement a set of sensors and actuators; and then, we implement a two-level hierarchical finite state machine to select an action in the next moment; at last, XML-based configuration for creating and setting up agent automatically is presented.
|