Publikation
Exploring Superior Structural Materials Using Multi-Objective Optimization and Formal Techniques
Rolf Drechsler; Stephan Eggersglüß; Nils Ellendt; Sebastian Huhn; Lutz Mädler
In: 6th IEEE International Symposium on Embedded Computing & System Design (ISED). International Symposium on Electronic System Design (ISED-06), December 15-17, Patna, India, 2016.
Zusammenfassung
The continuous progress in engineering requires increasingly powerful materials being utilized in complex constructions. Commonly used methodologies are hardly capable to explore materials with improved properties, i.e., fulfilling a certain performance profile. An alternative high throughput screening approach exists which allows to process a high number of samples while operating on micro scale. For structural materials, it is not possible to project material properties from the micro to the macro scale. Due to the complex nexuses between the process parameters (applied for sample synthesis), the high-dimensional space of the screening data and the uncertainties concerning material properties while scaling, conventional algorithms are not capable to cope with these data. Thus, this work proposes a general data flow which orchestrates multi-objective optimization approaches as well as formal techniques to handle this challenging computational task. In the end, a framework is drawn that determines the resulting material properties on a macro level by using screening data of micro samples and compares them against a performance profile. For the case that only a slight correlation exists, the framework proposes a batch of alternate process parameters which are, highly probable, leading to superior structural material.