Dynamic Metrics for Multi-agent Systems Using Aspect-Oriented Programming Application to DIMA Platform

摘要

The measurement is an important field in the software engineering. This field studies the assignment of a value to a software attribute using metrics. In fact, the use of metrics, which are the defined measurement methods and the measurement scales, in software engineering allows among other goals to control the quality of the software products, the quality of the software processes or the evolution of the software projects. Usually, we distinct two kinds of metrics: the static and the dynamic ones. The static metrics can be obtained from the source code of the software. On the other side, the dynamic metrics requires the execution of the software. Consequently, the static metrics are the most applied because of their relative simplicity compared to the dynamic metrics. However, we think that the dynamic metrics are more appropriate in some software categories. Multi-Agent Systems (MAS) is one of these categories because of the flexible nature of the agents. Consequently, we propose in this paper the using of the Aspect-Oriented Programming (AOP) to measure the dynamic metrics of multi-agent systems. In fact, the proposed approach provides several advantages like the simplicity, the reusability and the extensibility. In order to validate the proposed approach, we developed a tool that allows measuring of some metrics of DIMA-based applications.