Scalable BDI-based Multi-Agent System for Digital Design Reviews

Abstract

The increasing complexity in product development and the lack of knowledge exchange, for example, between development and manufacturing lead to unnecessary iteration loops and high costs. To overcome this situation, an option is the execution of a design review using multi-agent systems to provide designers with a digital assistance system for checking their CAD (computer-aided design) models regarding manufacturability using a milling process. This paper explores how to increase the scalability and robustness of multi-agent systems (MAS) for manufacturability assessment by using runtime-generated BDI (belief-desire-intention) agents and graph-based feature recognition. The applicability and validation of the presented approach is carried out by evaluating different milled part designs.