John Ford
Research Interests
My main interest lies in the development of faster and more
efficient algorithms for solving unconstrained optimisation
problems, although many of the ideas also have applications in other
areas. I focus particular attention on developing methods which are
able to combine old and new information in an effective manner.
Methods of this type currently under investigation are exhibiting
substantially improved performance when compared with "industry
standard" methods, particularly as the number of variables, and
therefore the difficulty of the problem, increases. I also work with
Edward Tsang in the area of Constraint Programming. My subsidiary
interests include chaos, with particular reference to the use of
graphics in studying the behaviour of nonlinearequation solvers,
and numerical algorithms on parallel computer systems.
Selected Publications
 Tsang, E. P. K., Ford, J. A., Lau, T. L., Mills, P., and
Voudouris, C., 'Operations research meets constraint programming:
some achievements so far', Proceedings, 15th National Conference of
the Australian Society for Operations Research (ASOR'99), Gold
Coast, Queensland (1999) pp. 13131324 [C4]
 Ford, J. A., 'Implicit Updates in Multistep QuasiNewton Methods',
Computers and Mathematics with Applications, 42: (2001) pp.
10831091 [C11]
 Ford, J. A. and Ghandhari, R. A., 'On the Use of Functionvalues in
Unconstrained Optimisation', Journal of Computational and Applied
Mathematics, 28: (1989) pp. 187198 [C11]
 Ford, J. A. and Saadallah, A. F., 'A rational function model for
unconstrained optimisation', Numerical Methods (Colloquia of Janos
Bolyai Mathematical Society), (1988) pp. 539563 [C4]
 Ford, J. A., 'Improved IllinoisType Methods for the Solutions of
Nonlinear Equations', Scientia Iranica, 4: (1997) pp. 2834 [C11]
 Ford, J. A. and Moghrabi, I. A., 'Minimum Curvature Multistep
QuasiNewton Methods', Computers and Mathematics with Applications,
31: (1996) pp. 179186 [C11]
 Ford, J. A. and Moghrabi, I. A., 'Multistep quasiNewton methods
for optimization', Journal of Computational and Applied Mathematics,
50: (1994) pp. 305323 [C11]
 Ford, J. A. and Moghrabi, I. A., 'Alternative parameter choices for
multistep quasiNewton methods', Optimisation Methods and Software,
2: (1993) pp. 357370 [C11]
 Ford, J. A., 'A Generalization of the JenkinsTraub Method',
Mathematics of Computation, 31: (1977) pp. 204213 [C11]
