Stochastic Optimal Control via Measure Relaxations
This work addresses the problem of scaling optimal control for stochastic systems, which is incremental as it builds on existing measure relaxation techniques.
The paper tackles the scalability challenge of optimal control for stochastic systems by reformulating it as a convex optimization over occupation measures, demonstrating the method on synthetic and real-world scenarios with cost functions learned from data via Christoffel polynomials.
The optimal control problem of stochastic systems is commonly solved via robust or scenario-based optimization methods, which are both challenging to scale to long optimization horizons. We cast the optimal control problem of a stochastic system as a convex optimization problem over occupation measures. We demonstrate our method on a set of synthetic and real-world scenarios, learning cost functions from data via Christoffel polynomials. The code for our experiments is available at https://github.com/ebuehrle/dpoc.