Huang, H. Y.2016-04-212016-04-211972Chambliss, Joe Preston. "Numerical experiments on the methods of dual matrices for function minimization." (1972) Master’s Thesis, Rice University. <a href="https://hdl.handle.net/1911/89048">https://hdl.handle.net/1911/89048</a>.https://hdl.handle.net/1911/89048Four algorithms of dual matrices for function minimization introduced in Ref. 1 are tested through several numerical examples. Three quadratic functions and five nonquadratic functions are investigated. For quadratic functions, the results show that the convergence is achieved in at most n+1 iterations, where n is the number of variables. Since one-dimensional search is not needed in these algorithms the total number of gradient evaluations for convergence is at most n+2. This represents a saving on the gradient evaluations versus 2n+1 required by the conventional quadratically convergent algorithms. For nonquadratic functions, the results show that these algorithms are very stable and efficient. Also, the effects of stepsize factor on these algorithms are investigated.36 ppengCopyright is held by the author, unless otherwise indicated. Permission to reuse, publish, or reproduce the work beyond the bounds of fair use or other exemptions to copyright law must be obtained from the copyright holder.ThesisRICE0083reformatted digitalThesis M.E. 1972 CHAMBLISS