Network configuration design includes complex tasks that are traditionally performed by domain experts. These efforts are time-consuming, costly and typically undocumented. In this work, we propose a new and innovative approach to dealing with network configuration design tasks. Our approach relies on artificial intelligence tools that utilize advanced knowledge representation, multistrategy learning, and problem solving methods. The approach consists of the development of an intelligent agent that assists human experts to solve network configuration design problems.; Throughout our effort, we studied various approaches to configuration design, formulated network configuration as a configuration design problem, specified an intelligent learning agent for network configuration design, utilized a multistrategy learning approach to develop a knowledge base for the intelligent agent which included defining an object ontology and learning rules for network configuration design, and formulated future directions to build a practical intelligent agent. |