Abstract
There are well established rival theories about the economy. These have, in turn, led to the development of rival models purporting to represent the economic system. The models are large systems of discrete-time nonlinear dynamic equations. Observed data of the real system does not, in general, provide sufficient information for statistical methods to invalidate all but one of the rival models. In such a circumstance, there is uncertainty about which model to use in the formulation of policy. Prudent policy design would suggest that a model-based policy should take into account all the rival models. This is achieved as a pooling of the models. The pooling that yields the policy which is robust to model choice is formulated as a constrained min-max problem. The minimization is over the decision variables and the maximization is over the rival models. Only equality constraints are considered.
A successive quadratic programming algorithm is discussed for the solution of the min-max problem. The algorithm uses a stepsize strategy based on a differentiable penalty function for the constraints. Two alternative quadratic subproblems can be used. One is a quadratic min-max and the other a quadratic programming problem. The objective function of either subproblem includes a linear term which is dependent on the penalty function. The penalty parameter is determined at every iteration, using a strategy that ensures a descent property as well as the boundedness of the penalty term. The boundedness follows since the strategy is always satisfied for finite values of the parameter which needs to be increased a finite number of times.
The global and local convergence of the algorithm is established. The conditions, involving projected Hessian approximations, are discussed under which the algorithm achieves unit stepsizes and subsequently Q-superlinear convergence.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
K.J. Arrow, F.J. Gould and S.M. Howe, “General saddle point result for constrained optimization,”Mathematical Programming 5 (1973) 225–234.
K.J. Arrow and R.M. Solow, “Gradient methods for constrained maxima, with weakened assumptions,” in: K. Arrow, L. Hurwicz and H. Uzawa, eds.,Studies in Linear and Nonlinear Programming (Stanford University Press, Stanford, CA, 1958) pp. 166–176.
R.G. Becker, B. Dwolatzky, E. Karakitsos and B. Rustem, “The simultaneous use of rival models in policy optimization,”The Economic Journal 96 (1986) 425–448.
M.C.B. Biggs, “The development of a class of constrained minimization algorithms and their application to the problem of power scheduling,” Ph.D. Thesis, University of London (London, England, 1974).
P.T. Boggs, J.W. Tolle and P. Wang, “On the local convergence of quasi-Newton methods for constrained optimization,”SIAM Journal on Control and Optimization 20 (1982) 161–171.
G.C. Chow, “Effective use of econometric models in macroeconomic policy formulation,” in: S. Holly, B. Rustem M. Zarrop, eds.,Optimal Control for Econometric Models (Macmillan, London, 1979) pp. 31–39.
C. Charalambous and J.W. Bandler, “Nonlinear minimax optimization as a sequence of leastpth optimization with finite values ofp,”International Journal of System Science 7 (1976) 377–391.
C. Charalambous and A.R. Conn, “An efficient algorithm to solve the min-max problem directly,”SIAM Journal on Numerical Analysis 15 (1978) 162–187.
G. Cohen, “An algorithm for convex constrained minimax optimization based on duality,”Applied Mathematics and Optimization 7 (1981) 347–372.
T.F. Coleman, “A note on ‘New Algorithms for constrained minimax optimization’,”Mathematical Programming 15 (1978) 239–242.
A.R. Conn, “An efficient second order method to solve the constrained minmax problem,” Report, Department of Combinatorics and Optimization, University of Waterloo (Waterloo, Ontario, 1979).
A.R. Conn and Y. Li, “The computational structure and characterization of nonlinear discrete Chebyshev problems,” Technical Report CS-88-39, Department of Computer Science, University of Waterloo (Waterloo, Ontario, 1988).
V.F. Demyanov and V.N. Malomezov,Introduction to Minmax (Wiley, New York, 1974).
V.F. Demyanov and A.B. Pevnyi, “Some estimates in minmax problems,”Kibernetika 1 (1972) 107–112.
S.R.K. Dutta and M. Vidyasagar, “New algorithms for constrained minmax optimization,”Mathematical Programming 13 (1977) 140–155.
J. Fuhrer and J. Haltmaier, “Minimum variance pooling of forecasts at different levels of aggregation,” Special Studies Paper 208, Federal Reserve Board (Washington, DC, 1986).
C. Granger and P. Newbold,Forecasting Economic Time Series (Academic Press, New York, 1977).
J.H. Hald and K. Madsen, “Combined LP and quasi-Newton methods for minimax optimization,”Mathematical Programming 20 (1981) 49–62.
S.P. Han, “Superlinear convergence of a minimax method,” Technical Report 78-336, Department of Computer Science, Cornell University (Ithaca, NY, 1978).
S.P. Han, “Variable metric methods for minimizing a class of nondifferentiable functions,”Mathematical Programming 20 (1981) 1–13.
M.J. Lawrence, R.H. Edmunson and M.J. O'Connor, “The accuracy of combining judgemental and statistical forecasts,”Management Science 32 (1986) 1521–1532.
S. Makridakis and R. Winkler, “Averages of forecasts: some empirical results,”Management Science, 29 (1983) 987–996.
J. Medanic and M. Andjelic, “On a class of differential games without saddle-point solutions,”Journal of Optimization Theory and Applications 8 (1971) 413–430.
J. Medanic and M. Andjelic, “Minmax solution of the multiple target problem,”IEEE Transactions on Automatic Control AC-17 (1972) 597–604.
W. Murray and M.L. Overton, “A projected Lagrangian algorithm for nonlinear minmax optimization,”SIAM Journal on Scientific and Statistical Computing 1 (1980) 345–370.
E. Polak and A.L. Tits, “A globally convergent, implementable multiplier method with automatic penalty limitation,”Applied Mathematics and Optimization 6 (1981) 335–360.
E. Polak, D.Q. Mayne and J.E. Higgins, “A superlinearly convergent min-max algorithm for min-max problems,” Memorandum No: UCB/ERL M86/103, Department of Electrical Engineering, University of California (Berkeley, CA, 1988).
M.J.D. Powell, “Variable metric methods for constrained optimization,” in: A. Bachem, M. Grötschel, B. Korte, eds.,Mathematical Programming: The State of the Art (Springer, Berlin-Heidelberg, 1983) pp. 288–311.
R.T. Rockafeller, “A dual approach to solving nonlinear programming problems by unconstrained minimization, “Mathematical Programming 5 (1973) 354–373.
B. Rustem,Projection Methods in Constrained Optimization and Applications to Optimal Policy Decisions (Springer, Berlin, 1981).
B. Rustem, “Convergent stepsizes for constrained optimization algorithms,”Journal of Optimization Theory and Applications 49 (1986) 136–160.
B. Rustem, “Methods for the simultaneous use of multiple models in optimal policy design,” in: C. Carraro and D. Sartore, eds.,Developments in Control Theory for Economic Analysis (Nijhoff-Kluwer, Dordrecht, 1987) pp. 157–186.
B. Rustem, “Equality and inequality constrained optimization algorithms with convergent stepsizes,” PROPE Discussion Paper 70, Department of Computing, Imperial College (London, 1989a).
B. Rustem, “A superlinearly convergent constrained min-max algorithm,” PROPE Discussion Paper 98, Imperial College (London, 1989b).
R.A. Tapia, “On secant updates for use in general constrained optimization,” Technical Report 84-3, Department of Mathematical Sciences, Rice University (Houston, TX, 1986).
R.S. Womersley and R. Fletcher, “An algorithm for composite nonsmooth optimization problems,”Journal of Optimization Theory and Applications 48 (1986) 493–523.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Rustem, B. A constrained min‐max algorithm for rival models of the same economic system. Mathematical Programming 53, 279–295 (1992). https://doi.org/10.1007/BF01585707
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF01585707