Application of the method of convexification to find a multivalued integral (Q1974994)

Computing the integral sums to approximate a multivalued integral requires the repeated construction of the convex hull of a positive-homogeneous function. An algorithm for numerical convexification of a positive-homogeneous function of two variables was proposed by \textit{D. B. Silin} and \textit{N. G. Trin'ko} [Zh. Vuchisl. Mat. Mat. Fiz., 34, No. 4, 545-548 (1994; Zbl 0818.65006)]. In this paper the above problem is considered for a positive-homogeneous function of \(n\) variables, with \(n>2\). The method of abbreviated convolution to construct the convex hull of a finite set of points is used. Applications of computing multivalued integrals are also discussed: constructing a pursuit strategy when solving a linear differential game and finding a viscosity solution of the Cauchy problem for a Hamilton-Jacobi equation.