A Framework for Implementing Deliberative Agents in Computer Games

Spyke3D: A new computer games oriented BDI Agent Framework

“…• Client-side. Agents are considered as separated applications of a client, which communicates with the game engine through asynchronous messages: Gamebots [8], Flexbot [9], Quakebot [10] and Pogamut [11], all mainly born for research reasons, are shining examples of this approach. A further game design effort is also needed and it often represents a serious constraint: according to Rabin [12] and Schwab [13], AI is usually added at the end of the game development and, at this level, the introduction of agent oriented programming may even bring to worse results.…”

“…Davies and Mehdi [8], Patel and Hexmoor [15], Weber et al [16] modeled BDI, or conceptually similar paradigm, bots for respectively Unreal Tournament [17], Counter-Strike [18] and StarCraft [19]: in each case BDI agents made bots more believable, providing a generic more realistic feeling to the game. However, some restrictions arise from this approach: BDI paradigm itself is not able to fully reproduce human-like cognitive processes [4]; moreover, as stated by Davies and Mehdi [8], adding a different bot concept on an existing game at a later stage is not an ideal solution. This process must, instead, occur concurrently with the game development, in order to better evaluate the agent behaviours.…”

Designing BOTs with BDI agents

“…al. researched creating a human-like AI, based on BDI paradigm, and design of framework for implementing deliberative agents in computer games [11] [12]. To test the behavior of AI they proposed an architecture using JACK [14] and linked it to Unreal Tournament [17] game engine via use of GameBots/JavaBot[18] technology which they used in both of their works [11] [12].…”

“…Even though, our simulation will be small as compared to real game Counter-Strike, the usage of Algorithm 1 and 2 to prove its effectiveness can be demonstrated. The reason is there already are related work ( [10], [11], [12], and [13]) that have demonstrated the integration of JAVA and Game Engines to implement BDI agent for bots in games, to proves that our miniature simulation of the game can be easily extended for use with current Game Engines in future. Henceforth, with this research we demonstrate the use of enhanced BDI model to fit the needs of bots in games and make them more human-like, believable, and to provide more realistic feel to the game.…”