A Compact Algorithm for Computing the Stationary Point of a Quadratic Function Subject to Linear Constraints.
Abstract
An algorithm is presented for computing the stationary point of a quadratic function of n variables subject to a set of m(m < or = n) linear equality constraints. The procedure is compact in the sense that it requires no two-dimensional arrays of computer storage beyond that needed to store the problem data. The use of a Householder orthogonal decomposition by the method should not degrade the numerical conditioning of the original problem. The method is applicable to problems with singular Hessian matrices, and can be adapted for use in a general quadratic programming algorithm. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 08, 1978
- Accession Number
- ADA058455
Entities
People
- J. T. Betts
Organizations
- The Aerospace Corporation