Architecture and collaborations among agents in mobile educational game

We describe a multi-agent based context-aware mobile educational game that can generate a series of learning activities for users doing on-the-job training in their working environment. In this paper, we reveal the multi-agent architecture (MAA) that is designed in the mobile educational game. The multi-agent architecture design enables development of a lightweight, flexible, and scalable game on the platform with limited resources such as mobile phones. Six agents have been implemented that worked together in this game. Each agent has its own tasks and sometimes needs help from other agents. Using agents in developing system can not only solve the limited resource issue, but also provide us a way to add further functions and to replace agents into the game easily to improve the game anytime. This paper focuses on designing the tasks that each agent needs to do and the collaborations that may happen among agents.