Computational design synthesis for conceptual design of robotic assembly cells

Abstract

Design synthesis is a fundamental engineering task that encompasses mapping a functional specification to a set of physical components and their topological relationships, where the mappings are typically non-unique. Even for moderately complex systems the consideration of all possible mappings is typically not feasible, making a manual exploration of the full design space impossible. In practice, this limitation often leads to sub-optimal designs. In this paper, we introduce a computational approach to design synthesis in conceptual design. Our approach is based around representing a design problem formally using SysML, and transforming this representation to a mixed-integer linear program. The generated mathematical optimization problem is then solved, and the mathematically optimal solution is transformed back to a SysML-based representation. We demonstrate our approach using the conceptual design of a robotic manufacturing cell as an example.