We present a decision-support framework for managing the charge and discharge operations in Battery Management Systems. Our goal-seeking system coordinates the charge and discharge operations of individual batteries in a collection without making any assumptions about the future behavior of the application. Our system allows users to implement different policies for scheduling the charge and discharge operations simply by changing the reward function to achieve different goals.
By reusing standard models for batteries, we demonstrate how we superpose a simple damage model that can incorporate well-known operational heuristics. This damage model allows users to design policies that must account for one or more damaged batteries in the collection. In the future, this approach can be realized on a networked embedded platform to address energy storage needs in a variety of applications.