EECE Department Building

Electronics Lab

Laser Lab

Hammam Lab

Home | Research | M.Sc. And Ph.D Thesis | Hybrid Model Based - Genetic Algorithm Diagnosis System

Hybrid Model Based - Genetic Algorithm Diagnosis System

Thesis Title: 
Hybrid Model Based - Genetic Algorithm Diagnosis System
Amr Mohamed Nour El-dean
Date of Birth: 
Thu, 09/09/1971
Previous Degrees: 
B.Sc. (EPM) 1993 - MTC
Registration Date: 
Wed, 01/10/2003
Awarding Date: 
Tue, 13/01/2009
External Supervisors: 

Dr. Mohamed, A. H.


Dr. Saad, E. M.
Dr. Shousha, A. M.
Dr. Nassar, A. M.

Key Words: 

Artificial intelligence (AI), Model based diagnosis (MBD), Model
based reasoning (MBR), Genetic algorithm (GA), Measurement
request list (MRL)


Nowadays artificial intelligence (AI) approaches are widely used in automating
the diagnostic process. Each approach has its power and weakness points.
However, integrating AI approaches can get the power of these approaches and
minimize the weakness. In this manner, this research introduces a diagnostic
system that integrate the model based diagnosis (MBD) and the Genetic
Algorithm (GA). GA is used in the proper cases to optimize the measurement
required list (MRL). Cases to apply GA are precisely determined by comparing
the MRL size and the count of fitness computation during GA. If MRL size is
smaller we use traditional deterministic technique to get the best MRL sequence,
while GA is applied when MRL size is bigger. Therefore, the proposed system can
decrease the cost of the measurements required for the diagnostic process. Also, it
can decrease the time, and complexity of the MBD systems. The proposed system
can be applied for diagnosis of many systems in different fields. It has been
applied for diagnosing complex electronic boards. The results obtained from the
proposed system have proved the improvement that had been done to the
performance of traditional MBD by using the GA optimization methodology