Early Validation and Global Optimization for Parametric Manufacturing Systems

Jasper Denkers

Creating software for manufacturing systems involves an interplay between mechanics, embedded software, and control software. Validating such software is costly because iterations require work from the different disciplines. The software is also complex, because the systems are parametric and we are interested in maximizing performance or minimizing resource waste. I’ve been working on a domain-specific language for specifying parametric manufacturing systems. With this DSL, we try to improve the software engineering process on two dimensions. First, the DSL introduces early validation by enabling mechanical engineers to carry out validation on the software specification level. Second, the DSL supports global optimization by generating constraint models for which optimal solutions are translated back to the specification level. In this talk I’ll present the DSL, the language engineering techniques behind it, and I’ll discuss my plans for evaluating the DSL within the context of Canon Production Printing. At Canon we have filed a patent for the application of this work to the printing domain. Therefore, I cannot talk about the application to printing, but I will talk about the generic language engineering approach and its application to an alternative, artificial, domain.

---

Previous: Robbert Krebbers | Proving Semantic Type Soundness in Iris
Next: Eelco Visser | A Direct Semantics of Disambiguation by Associativity and Priority Rules