Research Associate in "Mathematical optimal control and reinforcement learning for particulate systems" (m/w/d)
Research Framework
There is an ever increasing demand for the use of powder particles with micron size, providing very high specific surface area, in order to increase the dissolution rate. This is important to powder processing industries including pharmaceuticals when achieving the maximum effect of the products. However, handling and processing of such submicron size powders is challenging. Granulation is a process to modify the particle size distribution for improved material handling and dosing properties in powder processing. On the other hand, data-driven control, including reinforcement learning (RL) has been experiencing a renaissance as data availability and deep neural network (DNN) are rapidly progressing by pushes arising in multi-domains. In process engineering, challenges arise as data availability is rather limited and afflicted with relatively large errors. Thus, the use of data-driven or combined methods with classical approaches (grey-box models) seems appropriate to handle the complex relationships between particle and product properties.
Task Description
This project aims to develop the key scientific understanding of how to overcome these challenges in the granulation process by utilising mathematical tools concerning modelling, dynamical analysis and control, as well as the experimental data for parameter identification. Thereby, the main task is to create a predictive tool based on population balance modelling (PBM) and approximate method of moment (AMOM) for the design of continuous wet granulation on a twin-screw machine. The control task is to optimize the process with respect to a desired particle size distribution of the granule product, which is a critical quality attribute for subsequent unit operations, such as tabletting. Hereby, the screw configuration and different process parameters are the main impact factors within this optimization problem, which have to be addressed primarily. To this end, classical control approaches of optimal control theory or emerging learning-based algorithms, such as reinforcement learning, as well as a suitable combination thereof may be invoked.
This research work shall be conducted in close cooperation with academic researchers who are specialized in granulation and experimental research.
The research compiles from the following list of tasks.
- Developing of finite dimensional mathematical models using AMOM in the form of continuous ODEs for various classes of PBMs
- Analytical approximations of various functions appearing in the PBM in terms of orthogonal polynomials
- Optimal control of granulation process to obtain desired particle size distribution using Pontryagin maximum principle and dynamic programming
- Development of a learning based control algorithm based on reinforcement learning (RL) and deep-neural networks (DNN)
- Numerical solution of PDE and ODE models using classical methods or DNN
Qualification
- Above average university degree in applied mathematics, process engineering or control engineering
- Knowledge of at least one programming language: Matlab, Python, C++ is expected
- Knowledge in dynamical systems
- Proficiency in English or / and German is essential
- Highly motivated, eager to work within a team or independently.
Application procedure and deadline:
- Applications must include the following elements (as a single PDF file):
- Cover letter with a brief description of why you want to pursue research studies, about what your academic interests are, and how they relate to your previous studies and future goals
- CV including your relevant professional experience and knowledge
- Copies of diplomas and grades from previous university studies
- Two references
- List of publications
Send an email with the required documents to the address: @mec-apps@mv.uni-kl.de. The application deadline is 28. February 2025.
Interested candidates are encouraged to apply promptly, as applications will be processed as received, and the position may be filled before the deadline.
For more information, visit our webpage: https://mv.rptu.de/fgs/mec/research
