Explicit solutions of Riccati equations appearing in differential games (Q5895243)

When noncooperative control problems are tackled, a game theoretic approach is necessary: each control agent tries to optimize his own cost function which lies in conflict with others. An equilibrium solution must be sought, and Nash strategy is a first nature choice. Due to this noncooperation, the optimization problem of various players is strongly coupled and necessary conditions for Nash strategy lead to complex two- point boundary value problems. The aim of this paper is to find an explicit solution of a set of coupled asymmetric Riccati type matrix differential equations. The solution of such equations are generally difficult to obtain due to the permanent coupling between the player's strategies. Numerical techniques are widely used to obtain approximate or series solutions. An iterative algorithm for solving such coupled Riccati systems has been recently proposed by the author and \textit{H. Abou-Kandi} [SIAM J. Math. Anal. 19, No.6, 1425- 1430 (1988; Zbl 0663.65076)]. For the very particular case where the state weighting matrices are proportional, an analytic solution has been given by \textit{H. Abou-Kandil} and \textit{P. Bertrand} [Int. J. Control. 43, 997-1002 (1986; Zbl 0593.90092)]. We show that by means of an algebraic transformation we can decouple the player's optimization problems, leading in many cases to an explicit closed form solution of the game.