Abstract
The problem of reducing the state trajectory sensitivity to system parameter variations in linear-quadratic two-player differential games is considered. Open-loop and feedback Nash solutions are treated and two sensitivity reduction methods are developed for both solutions. The methods are based on a formulation where additional quadratic sensitivity terms are included in the players' cost functionals. The effectivity of these sensitivity reduction methods is compared by analysing the results in a simple numerical example.
Notes
† This work was supported in part by the Emil Aaltonen Foundation and the Yrjö Jahnsson Foundation.