Dynamic programming and maximum principle for discrete Goursat systems (Q1190320)

The author studies the boundary-value problem, obtained by discretization from the Goursat problem for the equation \(u_{xy}=g(x,y,u,a)\) where \(a\) is a control function. Using dynamic programming arguments he obtains conditions which are necessary and sufficient for \(a\) to be optimal. It allows him to write down explicit formulas for the exact or approximate solution.