Research Article

Agent-Oriented Software Engineering Methodologies: Analysis and Future Directions

Table 2

Comparative analysis of PASSI and O-MaSE methodologies.

MethodologyStrengthsLimitationsApplication domain

PASSI(i) Suggests a complete lifecycle methodology from requirement to code methodology(i) The need to refer simultaneously to various models in order to understand the system and the way it works and changes over time is a critical issue(i) A domain analysis approach for MASs product lines [58]
(ii) Integrates design models and concepts from both object-oriented (OO) and MAS using UML notation(ii) Every model offers its own set of notation and special concepts, resulting in an abnormal complexity in terms of vocabulary(ii) The development of a multiagent-based middleware for RFID asset management system using the PASSI methodology [59]
(iii) Refers to the most diffused standards: UML, the foundation for intelligent physical agents (FIPA), JAVA, and Rational Rose [32](iii) It does not support the environment model [32](iii) Patterns reuse in the PASSI methodology [32]

O-MaSE(i) Is comprehensive for building MAS(i) Some of the software applications are closed(i) Agent-based mixed-initiative collaboration project [60]
(ii) Is step-by-step lifecycle methodology MAS(ii) Management, product distribution, and testing and assessment have been absent(ii) O-MaSE: a customizable approach to developing multiagent development processes [61]
(iii) Provides guidance throughout the entire software development lifecycle(iii) There is the only one-to-one connection among agents in the system(iii) Developing a multiagent conference management system using the O-MaSE process framework [42]
(iv) Open systems are considered, thus agents can be created, deleted, or moved during implementation(iv) It does not clearly define the usage case model [23].ā€‰