Optimality and recursive feasibility guarantees for control algorithms

Establish theoretical guarantees for optimality and recursive feasibility of the conformal-prediction-based control algorithms proposed in the paper, including conditions under which the computed control sequences are optimal with respect to given objectives and remain feasible over time.

Background

Earlier sections present feasible control synthesis methods with probabilistic safety via statistical abstractions, but comprehensive optimality and recursive feasibility guarantees are not provided.

The authors explicitly highlight these guarantees as open technical questions that are important for practical deployment and robustness of the approach.