Online Control with Adversarial Disturbances
This work addresses robust control under adversarial conditions, which is incremental as it generalizes prior models to include adversarial noise and convex costs.
The paper tackles the problem of controlling a linear dynamical system with adversarial disturbances, aiming to minimize regret compared to an optimal controller with hindsight knowledge. The main result is an efficient algorithm that achieves nearly tight regret bounds.
We study the control of a linear dynamical system with adversarial disturbances (as opposed to statistical noise). The objective we consider is one of regret: we desire an online control procedure that can do nearly as well as that of a procedure that has full knowledge of the disturbances in hindsight. Our main result is an efficient algorithm that provides nearly tight regret bounds for this problem. From a technical standpoint, this work generalizes upon previous work in two main aspects: our model allows for adversarial noise in the dynamics, and allows for general convex costs.