| Summary: | 碩士 === 國立中正大學 === 電機工程研究所 === 99 === A cooperative negotiation and problem solving method, ANC (Automated Negotiation and Case-based reasoning), is proposed in this thesis to provide a suitable smart home service for multiple user requirements. The motivation in this thesis is that it is hard to have the common service, when different users propose different requirements. Therefore, in ANC, cooperation negotiation is used to resolve conflicts for making the requirements consistent, and problem solving is used to provide the suitable service for a common solution according to the common requirements. There are five stages in ANC. Issue acquisition is the first stage to acquire user requirements from user experience as the initial information. The purpose of using user experience is to obtain issues closer to user requirements. The second stage is to detect the conflicting issues for reducing negotiation overhead. Issue decision is the third stage to decide which of the issues needed to be negotiated for reducing negotiation overhead. Automated negotiation is the fourth stage to come to a consensus for the common problem. Each of the issues in the common problem is a conflict and needed to be negotiated. The fifth stage is to solve the common problem with the common solution. The common solution is the suitable service provided to users. In the fifth stage, case-based reasoning is the method to solve the common problem according to the past successful cases. To make the ANC system provide personalization, attributes weight learning method has been integrated by taking the advantage of the constant learning ability of case-based reasoning to make the attributes weight in accordance with user requirements. In this thesis, there are three contributions. The first is a new method proposed to provide a suitable service to users in a smart home. The second is the implementation of multi-user automated negotiation. The third is a case structure designed for multi-domain and multi-user.
|
|---|
