An Online Modeling Language for the Low-code Development of Belief-Desire-Intention (BDI) Agents
Bildiriler > Tam Metin Bildiri · 2023
Yazarlar: Burak Çelik, Geylani KARDAŞ, Barış Tekin Tezel
2nd International Symposium Series on Graduate Researches (ISGR 2023), Türkiye, İzmir, Uluslararası, Sözlü Sunum, Tam Metin Bildiri, 2023-2023, ss 186-196
Özet
<p>With their autonomy capabilities, agent-based systems have gained prominence in various industrial domains. Belief-Desire-Intention (BDI) architecture, a cornerstone of Agent-Oriented Software Engineering (AOSE), is commonly used to create proactive agents that interact effectively with their environment. Various domain-specific modeling languages (DSMLs) exist for the model-driven development of multi-agent systems (MAS) with a BDI approach. However, they are mostly dependent on the desktop libraries hindering widespread accessibility and online development. In this paper, we introduce a modeling language, called SAML, which enables the creation of BDI agents online. Our goal is to eliminate certain dependencies to the local computing resources and paving the way for a development environment that offers the low-code platform support in conjunction with a Software as a Service (SAAS) infrastructure. Additionally, the platform of SAML, also created in this study, servesas a platform-independent online modeling and development environment. The main objective is to promote the model-driven development of BDI agents using the low-code development techniques, one of the current trends in the software engineering field. By eliminating local dependencies, SAML can become a widely adopted tool for BDI agent development, fostering innovation and agility across various MAS applications. The paper presents a structured approach to achieve these goals, including research on the foundations of SAML, development of user interfaces for upper-level modeling, model-to-model transformation with JSON, and code transformation using JASON standards.</p>