Using Domain Specific Languages to Capture Design Synthesis Knowledge for Model-Based Systems Engineering

Design synthesis is a fundamental engineering task that involves the generation of a structural specification from a desired functional specification. Although the use of computer tools is common throughout the design process, design synthesis is often a task left to the designer. Formally capturing design synthesis knowledge in models and applying computational synthesis may result in better exploration of the design space and eliminate repetitive design tasks. In this paper, a graph-based framework for capturing and combining design synthesis knowledge is presented for scenarios involving the composition of well defined components into larger systems. This approach fits in the context of Model-Based Systems Engineering where a variety of formal models are used to represent knowledge about a system. This approach uses the Systems Modeling Language developed by The Object Management Group (OMG SysML™) to define both models of possible components and possible system architectures. The framework is illustrated by combining it with an evolutionary algorithm and applying it to an example problem of hydraulic circuit synthesis.