Model-Driven Engineering Approaches to Design Space Exploration
2013 (English)Report (Other academic)
During the design and deployment of increasingly complex distributed embedded systems, engineers are challenged by a plethora of design choices. This often results in infeasible or sub-optimal solutions. In industry and academia, general and domain-specific optimization techniques are developed to explore the tradeoffs within these design spaces, though these techniques are usually not adapted for use within a Model- Driven Engineering (MDE) process. In this paper we propose to encode metaheuristics in transformation models as a general design exploration method. This is complemented by an MDE framework for combining different heterogeneous techniques at different abstraction layers using model transformations. Our approach results in the seamless integration of design space exploration in the MDE process. The proposed methods are applied on the deployment of an automotive embedded system, yielding a set of Pareto-optimal solutions.
Place, publisher, year, edition, pages
2013. , 7 p.
IdentifiersURN: urn:nbn:se:kth:diva-118442OAI: oai:DiVA.org:kth-118442DiVA: diva2:606092
Report number SOCS-TR-2013.1. QC 201502102013-02-182013-02-182015-02-10Bibliographically approved