On diagonally preconditioning the truncated Newton method for super-scale linearly constrained nonlinear programming |
| |
Authors: | LF Escudero |
| |
Institution: | IBM Madrid Scientific Center, Paseo de la Castellana, 4-Madrid-1, Spain |
| |
Abstract: | We present an algorithm for super-scale linearly constrained nonlinear programming (LCNP) based on Newton's method. In large-scale programming solving the Newton equation at each iteration can be expensive and may not be justified when far from a local solution. For super-scale problems, the truncated Newton method (where an inaccurate solution is computed by using the conjugate-gradient method) is recommended; a diagonal BFGS preconditioning of the gradient is used, so that the number of iterations to solve the equation is reduced. The procedure for updating that preconditioning is described for LCNP when the set of active constraints or the partition of basic, superbasic and nonbasic (structural) variables have been changed. |
| |
Keywords: | Precontioning basic superbasic and nonbasic sets BFGS formula conjugate gradient truncated Newton direction de-activating strategy |
本文献已被 ScienceDirect 等数据库收录! |