Multi-agents for the Diagnostic Systems

Nowadays, critical and complex systems are widely used in our daily life. Modern technology provides great rapid changes in manufacturing these systems, and so continuous increasing in the complexity of their diagnostic systems. However, many intelligent systems and techniques have been developed in this manner. But, till now there is a great complexity appeared to have good performance for diagnosis these complex systems. This research introduces a novel diagnostic system that can use the multi-agents framework to simplify, extend and scale-up the diagnostic process. The proposed multi-agents diagnostic system can increase the accuracy of the diagnostic process by moving the tasks to the data rather than moving the data in the traditional diagnostic system. This can avoid the data losses and decrease the operation time. Also, the suggested system can simplify the diagnostic process by dividing the tasks of the system to be diagnosed into two main groups. Besides, this can help in increasing the accuracy and decreasing the time for achieving the required diagnosis. The proposed system has been applied for diagnosis the faults of three critical systems. They are: product handling mechanism system and a security network for a radiation centre, and a radiotherapy system. Its results are compared to three diagnostic systems. Their evaluation process proves that the suggested system has a significant improvement of the diagnostic systems' performance.