A COMPUTATIONALLY PRACTICAL INTERIOR-POINT TRUST-REGION ALGORITHM FOR SOLVING THE GENERAL NONLINEAR PROGRAMMING PROBLEMS

Abstract

An interior-point trust-region algorithm for solving the general nonlinear programming problem is proposed. In the algorithm, an interiorpoint Newton method with Coleman-Li scaling matrix is used. A trustregion globalization strategy is added to the algorithm to insure global convergence. A projected Hessian technique is used to simplify the trustregion subproblems.
A Matlab implementation of the algorithm was used and tested against some existing codes. In addition, four case studies were presented to test the performance of the proposed algorithm. The results showed that the algorithm out perform some existing methods in literature.