Abstract
The projection of the origin onto an n-dimensional polyhedron defined by a system of m inequalities is reduced to a sequence of projection problems onto a one-parameter family of shifts of a polyhedron with at most m + 1 vertices in n + 1 dimensions. The problem under study is transformed into the projection onto a convex polyhedral cone with m extreme rays, which considerably simplifies the solution to an equivalent problem and reduces it to a single projection operation. Numerical results obtained for random polyhedra of high dimensions are presented.
Similar content being viewed by others
References
E. A. Nurminskii, “Method of Successive Projections for Solving the Least Distance Problem for Simplexes,” Elektron. Zh. “Issledovano v Rossii” 160, 1732–1739 (2004); http://zhurnal.ape.relarn.ru/articles/2004/160.pdf.
E. A. Nurminskii, “Convergence of the Suitable Affine Subspace Method for Finding the Least Distance to a Simplex,” Zh. Vychisl. Mat. Mat. Fiz. 45, 1996–2004 (2005) [Comput. Math. Math. Phys. 45, 1915–1922 (2005)].
E. A. Nurminskii, “Acceleration of Iterative Methods of Protection onto a Polyhedron,” Dal’nevost. Mat. Sb., 1, 51–62 (1995).
C. Lemarechal and J.-B. Hiriart-Urruty, Convex Analysis and Minimization Algorithms II (Springer-Verlag, Berlin, 1993).
V. F. Dem’yanov and A. M. Rubinov, Fundamentals of Nonsmooth Analysis and Quasi-Differential Calculus (Nauka, Moscow, 1990) [in Russian].
M. A. Kozlov, S. P. Tarasov, and L. G. Khachiyan, “Polynomial Solvability of Convex Quadratic Programming,” Zh. Vychisl. Mat. Mat. Fiz. 20, 1319–1323 (1980).
Octave, http://www.octave.org.
ILOG CPLEX, http://www.ilog.com/products/cplex/.
Author information
Authors and Affiliations
Corresponding author
Additional information
Original Russian Text © E.A. Nurminski, 2008, published in Zhurnal Vychislitel’noi Matematiki i Matematicheskoi Fiziki, 2008, Vol. 48, No. 3, pp. 387–396.
Rights and permissions
About this article
Cite this article
Nurminski, E.A. Projection onto polyhedra in outer representation. Comput. Math. and Math. Phys. 48, 367–375 (2008). https://doi.org/10.1134/S0965542508030044
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1134/S0965542508030044