Model-Based Testing (MBT) aims to automate test case generation and execution from abstract models representing the behavior of the system under test (SUT). MBT stages parallelization and monitoring could improve the overall process performance in complex systems and resource shortages. Agent-based software testing (ABST) consists of intelligent agents applied to complex software testing tasks. ABST approaches could enhance software testing due to multi-agent systems autonomy, independence, parallel activation, and decision-making features. This study presents an agent-based tool for MBT case generation and execution. The tool comprises two components: the MBT component based on the MBT4J tool and a JADE multi-agent component. The multi-agent component implements a coordinator agent (CA), a monitor agent (MA), and several test case generation and execution agents (TGEA). The responsibilities of agents include test case planning, generation and execution, and results synthesis. The results suggest that low levels of TGEA achieve acceptable coverage metrics in straightforward models. Productivity provides the best results in the first execution cycles.
Tipo de publicación: Conference Paper
Publicado en: 2021 IEEE V Jornadas Costarricenses de Investigación en Computación e Informática (JoCICI)Autores
- Jose Ramírez-Méndez
- Christian Quesada-López
- Marcelo Jenkins
Proyecto asociado a la publicación