Agent Collaboration in a Multi-Agent-System for Analysis and Optimization of Mechanical Engineering Parts

Abstract In mechanical engineering, designers have to review a designed artefact iteratively with different domain experts, e.g. from manufacturing, to avoid later changes and find a robust, optimized design. To support the designer, knowledge-based engineering offers a set of approaches and techniques that formalize and implement engineering knowledge into generic product models or decision support systems. An implementation which satisfies especially the concurrent nature of today’s design processes and allow for multi-objective decision-making is multi-agent systems. Such systems consist of entities that are capable of autonomous action, interact intelligently with their environment, communicate and collaborate. In this paper, such a multi-agent system is discussed as extension for a computer-aided design software where the agents take the role of domain experts, like e.g. manufacturing technologists and make suggestions for the optimization of the design of mechanical engineering parts. A focal point is set on the collaboration concept of the single agents. Therefore, the paper proposes the use of an action-item-list as central information and knowledge sharing platform.