REDUCIBLE NONLINEAR PROGRAMMING PROBLEMS (SEPARABLE LEAST SQUARES)

[摘要] In this thesis we are concerned with general nonlinear programming problems in which the variables can be naturally separated into two groups. This separation has the property that if the variables in one of the groups are fixed, then it is easy to solve the problem in terms of the unspecified variables. The existing work in nonlinear least squares has been extended to handle inequality constraints and further extended to certain classes of nonlinear programming problems which have special structure.

[发布机构] Rice University