Access Type

Open Access Dissertation

Date of Award


Degree Type


Degree Name



Electrical and Computer Engineering

First Advisor

Hao Ying


HIV/AIDS is a global problem. Its treatment is dependent on the physician experts' opinion. A system which is capable of supporting the treatment decision will be desired. Recently, the HIV/AIDS treatment regimen selection system appeared in literature that utilized theory of fuzzy discrete event system (FDES) to capture the meaning of experts' knowledge; a form of consensus involving estimated points and type-1 fuzzy sets. The goal was to assign exact matching regimens as close as possible to those regimens preferred by the experts for patients. The system performance was 80% of satisfaction level with the 35 retrospective patients. Extracting experts' knowledge into the consensus forms would not be possible without being compromised by the experts. With equal respective experts, if one insists on his/her values, then the consensus would not be achieved. Conversely, the FDES theory would be no longer to handle such conflict. The theory of extended fuzzy discrete event system (EFDES) extended the FDES theory that type-2 fuzzy sets would be allowed to be used in the system. This dissertation is to apply the EFDES theory to the HIV/AIDS treatment regimen selection system. Seven scenarios of the diversity of experts' knowledge representation were categorized for the system. The MATLAB was implemented to model the system. Genetic algorithm in MATLAB's Direct Search Toolbox was used to search an optimal vector of 26 weights for system parameters regarding the experts' regimen-choices. As the same input of the retrospective patient data for the FDES-based system, the overall means of simulation results of EFDES-based system demonstrated the degree of matching regimens being 80%. That result would be the same performance level of the FDES-based system as well. The EFDES-based system performance with self-learning provided the overall satisfaction level of above 80%. Moreover, the EFDES-based system with use of the type-2 fuzzy set gained the benefit on the extraction of diverse and uncertainty experts' knowledge and expertise.