Testing HMAS-based applications: An ASPECS-based approach. (November 2015)

• Record Type: Journal Article
• Authors: Dehimi, Nour El Houda; Mokhati, Farid; Badri, Mourad
• Is Part Of: Engineering applications of artificial intelligence. Volume 46:Part A(2015:Oct.)
• Page Start: 232
• Page End: 257
• Publication Date: 2015-11
• Subjects: Holonic agent -- Model-based testing -- Genetic algorithm -- Quality assurance; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2015.09.013
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Organization-Based Multi-Agent Systems are receiving increasingly more attention in the field of complex and distributed systems development. Holonic Multi-Agent Systems (HMAS) are considered as well known organization allowing the construction of complex systems that are efficient in terms of use of resources, highly resilient to disturbances and adaptable to changes in their environment. However, HMAS-based applications development is not mature enough yet. Particularly, testing activities, which represent an important task in their quality assurance, are not well covered. This paper aims at proposing a new model-based testing technique for holonic agents. The technique uses genetic algorithms and takes into account the evolution (successive versions) of an agent. The approach is organized in two main phases that are conducted iteratively. The first phase focuses on the detection of a new version of an agent under test. The second phase addresses the testing of each new detected version. The new version of the agent is analyzed in order to generate a behavioral model on which is based the generation of test cases. The test cases generation process focuses on the new (and/or changed) parts of the agent behavior. In this way, the technique supports an incremental update of the test cases, which is a crucial issue. A software tool that we developed supports the proposed technique. The approach and associated tool are illustrated using a concrete case study.