Batteries, being the most valuable and defining component of electric vehicles (EVs), are of paramount importance to future mobility. The demand for battery testing facilities is soaring, with capacities being reserved years ahead. Here, prototype battery components (cells, modules) are the unit under test (UUT) to potentially thousands of test instruments (e.g., cyclers).
The planning process involves a multitude of tasks, including the organization of workloads, parameters, and test procedures, mapping of UUTs to test equipment, as well as the setup of equipment. The task of reliably planning and managing the usage, control, and inventory of thousands of test channels surpasses human capabilities due to its complexity. Classic optimization methods on the other hand reach their limits due to plant uncertainties (eg., unavailability due unplanned maintenance) and sometimes unpredictable behavior of the unit under test itself. In this work, new ML-based algorithms shall be implemented, analyzed, and compared to existing methods in managing test labs.

• Analyze and adapt data models for battery testing
• Implement simple test environment to simulate battery test lab
• Analyze state of the art in ML applicable to plant management 
• Select appropriate approaches and implement prototypes for tasks in the battery testing facility (e.g. allocate test equipment)
• Design an agent prototype that interacts with the environment
• Design dashboard prototypes to properly communicate necessary changes in resource allocation