Adaptive Learning of Assimilation and Accommodation for Intelligent Agent System
| Autor: | Kai-Chung Chang, 張凱鈞 |
|---|---|
| Rok vydání: | 2007 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 95 Adaptive learning is broadly used in agent learning. Cooperation learning is also a crucial ability in multi-agent system to improve the speed of convergence and quality of learning. A flexible also easily approach is needed for agents to learn individually and cooperatively. This paper presents an adaptive approach to address a kind of adaptive learning for intelligent agent. The agent can learns adaptively, and can communicate with other agents for cooperative learning by using this adaptive process. A knowledge processing cycle is proposed here for intelligent agent to mimic the human’s learning process. This process regards two important parts, the assimilation process for learning new information, the accommodation process for resolution when new information have some conflict with agent’s proper known. The cognitive structure is the representation of agent’s knowledge base which used on these two processes for learning, the structure here we fit it into BDI-agent model. We use the pursuit-evasion game to explain our approach which can explain multi-agent’s behavior easily, and for multi-agent system we use the JADE (Java Agent DEvelopment Framework) as our agent platform to handle agent’s essential supports in development. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|