Abstract:
The efficiency of multi-agent system (MAS) design mainly relies on the quality of a
conceptual architecture. Thereby, the quality substantially influences the software
system architecture, and plays a key role in the description of the initial architecture.
Hence, quality properties, such as understandability, complexity, readability,
testability, maintainability, reusability, etc should be considered at an early stage in
the software development process. It is worth noticing that, large systems such as
multi agents systems require many communications and interactions to fulfill their
tasks, which may leads to complexity of architecture design (AD).
This thesis attempts to clarify the complexity situations that might happen during the
description of architecture design through many aspects represented in the abstraction,
modularity, and modeling by introducing an approach that aims to put a set of
guidelines to minimize the effects of complexity, provides clarification for each
guideline and guides agent systems developers in order to design architectures with
high quality, low complexity and understandable.
The approach is applied on a case study to books recommendations system which
based on multi agent systems where the complexity is measured by complexity task
measurement (CTM) which based on use case point method. The solution has also
displayed the complexity results before and after applying the approach.
