Unified Approach to Unconstrained Minimization (I). Generation of Conjugate Directions for Unconstrained Minimization Without Derivatives (II).
CALIFORNIA UNIV BERKELEY DEPT OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCES
Pagination or Media Count:
Several important classes of algorithms for unconstrained minimization when applied to a quadratic function with Hessian A, may be regarded as being alternative ways to effect certain matrix factorizations of or with respect to A. This approach leads to a clear insight into the basic equivalence of many algorithms that are implemented in very different ways and which differ in their informational requirements. It also enables their presentation within a unified framework. In Part 2, the author analyzes in detail a particular technique for generating conjugate directions for unconstrained minimization without derivatives. Modified author abstract
- Theoretical Mathematics