English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

While $\mathcal{H}_\infty$ methods can introduce robustness againstworst-case perturbations, their nominal performance under conventionalstochastic disturbances is often drastically reduced. Though this fundamentaltradeoff between nominal performance and robustness is known to exist, it isnot well-characterized in quantitative terms. Toward addressing this issue, weborrow from the increasingly ubiquitous notion of adversarial training frommachine learning to construct a class of controllers which are optimized fordisturbances consisting of mixed stochastic and worst-case components. We findthat this problem admits a stationary optimal controller that has a simpleanalytic form closely related to suboptimal $\mathcal{H}_\infty$ solutions. Wethen provide a quantitative performance-robustness tradeoff analysis, in whichsystem-theoretic properties such as controllability and stability explicitlymanifest in an interpretable manner. This provides practitioners with generalguidance for determining how much robustness to incorporate based on a priorisystem knowledge. We empirically validate our results by comparing theperformance of our controller against standard baselines, and plotting tradeoffcurves.

Performance-Robustness Tradeoffs in Adversarially Robust Linear-Quadratic Control