A BDI Game Master Agent
2 likes753 views
The document discusses the development of a game master agent for computer role-playing games (RPGs) using a Belief-Desire-Intention (BDI) framework to enhance storytelling flexibility. The prototype was implemented in Neverwinter Nights and evaluated based on user experience, showing that players had a better gaming experience when playing with the agent. The approach aims to improve RPG flexibility and decrease game development time.
![A BDI Game Master Agent
for Computer Role-playing Games
Bao Luong, John Thangarajah, Fabio Zambetta, Mahmud Hasan
s3272297@student.rmit.edu.au, johnt@rmit.edu.au, fabio.zambetta@rmit.edu.au, s3317339@student.rmit.edu.au
1. Problem
Though computer Role-playing games
(RPGs) are based on classic RPGs, they
lack exible storytelling. Game developers
usually employ a prescripted quest system
to maintain the overall setting of the sto-
ryline. However, by doing so they also
restrict player's interactions.
2. Our Approach
We propose an approach to
• Develop a Game Master (GM) agent
using the BDI agent framework.
• Implement it into a computer RPG,
i.e. Neverwinter Nights (NWN).
• Evaluate the BDI GM agent.
3. Contributions
We introduced a technique that uses
BDI framework for developing a BDI
GM agent. It can help
• Improve the exibility of computer
RPGs and introduce better gaming
experience.
• Decrease game development time.
4. BDI Game Master Agent
We demonstrated our approach in a NWN game scenario where the GM directs players
into a guarded room. There are three main steps
Step 1: Identifying game-flow diagrams
The game-ow diagrams are constructed by mapping the scenario objectives
and their related solutions into operation steps (rectangle) and decision nodes
(diamond shape with nodes).
Step 2: Mapping Operation Steps to Goal Tree
For operation steps that are joined at a decision node, they become OR connected
sub-goals. For operation steps that are not, they become AND connected sub-goals.
Step 3: Constructing BDI Goal-Plan Tree
This step is divided into two minor steps
First. We create plans for each goal to design an initial goal-plan tree by
1. Expand a goal with alternative plans 2. Create a new plan
Second. We rene the initial goal-plan tree by
1. Add new plans to single-plan goals. 2. Add new sub-goals to plans.
3. Group similar goals and plans. 4. Collapse sub-goals into single plans.
5. Implementation
We developed a prototype GM agent
based on a scenario in NWN, and inte-
grated it into the engine through the Nev-
erwinter Nights eXtender plugin.
The communication layer between the agent and
the NWN engine.
6. Experimental Setup
Players played both versions of the
game scenario with and without the agent.
We then evaluated user experience
following 5 criteria that are
• Flexibility.
• Replayability.
• Clear objective setting.
• Interest.
• Future appeal.
7. Results
Overall, all participants (31 persons)
experienced better gaming experience af-
ter playing the scenario with the GM
agent.
Summarised results from the survey.
A = Game w/o GM, B = Game with GM.
1. Bratman, M. E. (1999). Intention, Plans,
and Practical Reason. Cambridge University
Press, Cambridge, England.
2. Padgham, L. and Winiko, M. (2004). Devel-
oping Intelligent Agent Systems: A Practi-
cal Guide, pages 107 to 135. John Wiley and
Sons.
3. BioWare Corp (2002). Neverwinter Nights.
[CD-ROM]. Computer role-playing game.](https://image.slidesharecdn.com/bdigm-poster-130429192128-phpapp01/85/A-BDI-Game-Master-Agent-1-320.jpg)
A BDI Game Master Agent
- 1. A BDI Game Master Agent for Computer Role-playing Games Bao Luong, John Thangarajah, Fabio Zambetta, Mahmud Hasan [email protected], [email protected], [email protected], [email protected] 1. Problem Though computer Role-playing games (RPGs) are based on classic RPGs, they lack exible storytelling. Game developers usually employ a prescripted quest system to maintain the overall setting of the sto- ryline. However, by doing so they also restrict player's interactions. 2. Our Approach We propose an approach to • Develop a Game Master (GM) agent using the BDI agent framework. • Implement it into a computer RPG, i.e. Neverwinter Nights (NWN). • Evaluate the BDI GM agent. 3. Contributions We introduced a technique that uses BDI framework for developing a BDI GM agent. It can help • Improve the exibility of computer RPGs and introduce better gaming experience. • Decrease game development time. 4. BDI Game Master Agent We demonstrated our approach in a NWN game scenario where the GM directs players into a guarded room. There are three main steps Step 1: Identifying game-flow diagrams The game-ow diagrams are constructed by mapping the scenario objectives and their related solutions into operation steps (rectangle) and decision nodes (diamond shape with nodes). Step 2: Mapping Operation Steps to Goal Tree For operation steps that are joined at a decision node, they become OR connected sub-goals. For operation steps that are not, they become AND connected sub-goals. Step 3: Constructing BDI Goal-Plan Tree This step is divided into two minor steps First. We create plans for each goal to design an initial goal-plan tree by 1. Expand a goal with alternative plans 2. Create a new plan Second. We rene the initial goal-plan tree by 1. Add new plans to single-plan goals. 2. Add new sub-goals to plans. 3. Group similar goals and plans. 4. Collapse sub-goals into single plans. 5. Implementation We developed a prototype GM agent based on a scenario in NWN, and inte- grated it into the engine through the Nev- erwinter Nights eXtender plugin. The communication layer between the agent and the NWN engine. 6. Experimental Setup Players played both versions of the game scenario with and without the agent. We then evaluated user experience following 5 criteria that are • Flexibility. • Replayability. • Clear objective setting. • Interest. • Future appeal. 7. Results Overall, all participants (31 persons) experienced better gaming experience af- ter playing the scenario with the GM agent. Summarised results from the survey. A = Game w/o GM, B = Game with GM. 1. Bratman, M. E. (1999). Intention, Plans, and Practical Reason. Cambridge University Press, Cambridge, England. 2. Padgham, L. and Winiko, M. (2004). Devel- oping Intelligent Agent Systems: A Practi- cal Guide, pages 107 to 135. John Wiley and Sons. 3. BioWare Corp (2002). Neverwinter Nights. [CD-ROM]. Computer role-playing game.