Related Experiment Videos
Patterns of agent interaction scenarios as use case maps.
1Department of Math and Computer Science, California State University, Hayward, CA 94542, USA. billard@csuhayward.edu
Summary
This study introduces use case maps (UCMs) to visualize agent interaction scenarios. It adapts software design patterns to structure agent behavior for cleaner, more organized decentralized systems.
Area of Science:
- Computer Science
- Artificial Intelligence
- Software Engineering
Background:
- Autonomous agent interactions are complex and require structured representations.
- Traditional software design patterns offer insights into system architecture.
- Use Case Maps (UCMs) provide abstract system descriptions suitable for agent scenarios.
Purpose of the Study:
- To adapt object-oriented design patterns for describing autonomous agent interaction scenarios.
- To present these adapted patterns using Use Case Maps (UCMs).
- To demonstrate how UCMs can clarify agent behavior and system structure.
Main Methods:
- Identification of relevant "gang of four" design patterns applicable to agent interactions.
- Representation of these patterns as Use Case Maps (UCMs).
- Performance analysis of a selected UCM to illustrate benefits of early abstraction.
Main Results:
- Eight design patterns were identified as suitable for agent interaction architectures.
- Seven patterns were adapted to balance decentralized agent behavior with organized structure.
- UCMs effectively visualize causal paths in agent behavior derived from design patterns.
Conclusions:
- Use Case Maps (UCMs) offer a valuable method for abstractly representing autonomous agent interactions.
- Software design patterns can be effectively translated into UCMs to structure agent systems.
- This approach enhances the clarity and organization of complex agent behaviors.