Servicenavigation

Project Details

Back

Projekt Print View

Scalable Design Space Exploration via Answer Set Programming

Subject Area Computer Architecture, Embedded and Massively Parallel Systems
Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
Term from 2015 to 2023
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 269264143
 
Final Report Year 2024

Final Report Abstract

Embedded systems (ES) are application-specific computers that can be found in almost every product today. They usually have to meet real-time, power and space requirements, among others, while being reliable and cost-efficient. These often contradictory design goals can only be met because each ES is developed for a specific and therefore restricted group of applications. Examples of embedded computing systems can be found in medical devices, industrial automation systems, automotive and train control systems, IoT devices (Internet of Things), etc. Due to their constantly increasing computing requirements, ES are usually implemented as heterogeneous many-core computers, i.e. they consist of many different computing cores of different types. In addition to selecting the number and type of computing cores, decisions must also be made about the communication infrastructure, memory organization, task distribution, communication routing, and task and communication scheduling. All these decisions lead to incredibly large design spaces that need to be explored effectively. To summarize, increasing application complexity, coupled with increasingly complex computing platforms, makes it difficult to make good design decisions and thus optimize the end product. As a result, new tools and methods are needed to enable automatic and effective exploration of design options at system level. In this project, novel Design Space Exploration (DSE) methods based on Answer Set Programming (ASP), a declarative programming paradigm for combinatorial search problems, were developed and investigated. The project has made important contributions in several areas: (1) Overcoming bottlenecks in determining multi-hop communication routing by utilizing the reachability expression capabilities built into ASP. (2) A powerful DSE method based on ASPmT by tightly integrating the background theory into ASP, which enables both scheduling and multi-objective optimality filtering in the background theory. (3) Integration of generative product design into the ASPmT-based DSE, enabling faster and more targeted exploration of derived product variants. (4) Automatic platform generation during DSE to reduce the overall specification effort. All these contributions were only possible thanks to the close collaboration of two research groups from different areas, namely ASP solving and DSE at system level.

Publications

 
 

Additional Information

© 2025 DFG Disclaimer / Copyright / Privacy Policy

Textvergrößerung und Kontrastanpassung