SPHERICAL PROGRAMMING: A CONVEX PROGRAMMING ALGORITHM.
Abstract
A new algorithm is developed for solving the problem of maximizing a function, f(x), of n variables subject to m linear inequality constraints. The procedure consists of solving a sequence of subproblems which require the maximization of f(x) over a hypersphere. A simple algorithm is developed for solving the subproblems. The sequence of subproblem optima is shown to converge to the constrained optimum of f(x) if f(x) is concave. A discussion is given of our limited computational experience with the algorithm. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 08, 1968
- Accession Number
- AD0680463
Entities
People
- Herman Otto Hartley
- R. R. Hocking
- S. W. Mcguire
Organizations
- Texas A&M University