Reliability Prediction Approach for Multi-Agent Systems

Abstract

The software’s are created to implement the needs of the customer in reality, and this raises the need to measure the applicability and accuracy of the software. Software reliability engineering is based on models and measurements that quantify the software reliability.

In our thesis, we have proposed a reliability prediction method of Multi agent systems integrating metrics depending on the agent's behavior and MAS characteristics.

For the development of our approach we used machine learning regression algorithm that is based on finding the correlation between the proposed metrics in order to predict the agent’s reliability.