Abstract
The environment is an important but overlooked piece in the construction of multiagent-based scenarios. Richness, believability and variety of scenarios are inseparable from the environment because every action and interaction of agents is based around the environment they are situated in. The prerequisite, however, is that agents must be able to understand the environment and capture its dynamic nature. This paper proposes a cognitive middle layer between agent minds and the environment. Aspects of the reality are mapped to concepts in the middle layer, through which agents can feel and reason about the real environment. The middle layer is modelled with a structured specification based on Web Ontology Language (OWL) to be extensible and reusable. Environmental concepts are integrated into the goal processing of agents to trigger intentions. This paper also reports our initial investigation about the design of a simulation system for multiple environment-aware agents and multiple users.
| Original language | English |
|---|---|
| Pages (from-to) | 57-73 |
| Number of pages | 17 |
| Journal | Lecture Notes in Artificial Intelligence (Subseries of Lecture Notes in Computer Science) |
| Volume | 3374 |
| DOIs | |
| State | Published - 2005 |
| Externally published | Yes |
| Event | First International Workshop on Environments for Multi-Agent Systems - New York, NY, United States Duration: 19 07 2004 → 19 07 2004 |