本論文使用DAML (DARPA Agent Markup Language) 來描述某給定的領域知識的Ontology，以DAML ontology為基礎，我們可以建構一個目標導向個人代理人。本方法論分別藉由新增DAML ontology與non-agent系統中加入Wrapper，使整合agent與non-agent系統更為容易。將領域知識 (domain knowledge) 從操作知識(operated knowledge)中分開，agent加入系統時，ontology可以容易地被新增。因此，不同的Agent可以互相溝通，只要ontology已經更新了。更進一步，因為規則與限制被描述於DAML中，agent可以從DAML ontology中做推理，且有智慧地互動。另外，在達成目標時，如果有不可預期的情況發生，agent藉由DAML ontology作推理，做出不同的動作行為。 This thesis use DAML (DARPA Agent Markup Language) to describe the ontology of a given domain knowledge. Based on the DAML ontology, this goal-oriented personal agent can be built. This approach facilitates integrate of agents and non-agent systems, respectively, by updating DAML ontology and by adding wrapper. By separating domain knowledge (ontology) from operated knowledge, ontology can be easily updated, when an agent joins the system. Thus, different agents can communicate with each other as long as the ontology has updated. Furthermore, because that rules and constricts can be described in DAML, agents can infer from the DAML ontology and interact intelligently. In addition, in pursuing a goal, if unforeseen situation occurs, the agents can act differently by inferring form the DAML ontology.
