Project: Pareto Analysis for System-Level Simulations
Description
Deriving and visualizing the impact of parameter change to down-stream variables
We have models to run simulations of different systems and their sub-components, which allows to optimize configurations for the scanner to achieve a given objective.
Our software allows to run a multi-objective optimization and derive the ideal input conditions to achieve the desired outcome. Often, more than a single solution will result into the minimization of the targeted loss function. This results in a so called “Pareto-front”, i.e., a multi-dimensional surface along which the desired cost is optimal.
In this project, the student will investigate ways to obtain and visualize the Pareto-front. Minimal functionality is already in place for the extraction of the front, but more advanced methods can be investigated. Moreover, advanced visualization of the front will be key for the outcome of the project, as the front is expected to lay on a multi-dimensional space.
- (Advanced) Visualization knowledge
- (Basic) SW Engineering skills
- (Basic) Knowledge of PyTorch and ML
- Javascript visualization libraries (e.g., D3.js)
- React
- A method to extract the pareto front from the Virtual Scanner
- A set of visualizations in React.js to analyze the resulting Pareto front
The project will be supervised by Nicola Pezzotti at ASML Research and Stef van den Elzen and/or Anna Vilanova at TUe. The student is expected to be at ASML in Building 7 for 3 days a week
Details
- Supervisor
-
Stef van den Elzen
- Secondary supervisor
-
Anna Vilanova
- External location
- ASML
- Interested?
- Get in contact