Title: A Subspace Trust Region Algorithm


			       Yuan Ya-xiang
Institute of Computational Mathematics and Scientific/Engineering Computing
		  Academy of Mathematics and System Sciences
		          Chinese Academy of Sciences
			 P.O. Box 2719, Beijing 100080
			  People's Republic of China
			   Email: yyx@lsec.cc.ac.cn
		      Web page: http://lsec.cc.ac.cn/~yyx


Abstract:
A trust region algorithm for unconstrained optimization  is
constructed using subspace techniques. The new method can be
viewed as an improvement on the limited memory quasi-Newton method
with trust regions. Due to the special structures of the subspace
trust region subproblems, the method can be applied to very large
scale problems. Convergence properties of the method are analyzed.
Numerical results using the CUTE examples are reported and it
shows that the new algorithm is efficient.