A comparative study of several general convergence conditions for algorithms modeled by point-to-set maps
A general structure is established that allows the comparison of various conditions that are sufficient for convergence of algorithms that can be modeled as the recursive application of a point-to-set map. This structure is used to compare several earlier sufficient conditions as well as three new sets of sufficient conditions. One of the new sets of conditions is shown to be the most general in that all other sets of conditions imply this new set. This new set of conditions is also extended to the case where the point-to-set map can change from iteration to iteration.
Key wordsOptimization Algorithms Convergence Conditions Point-to-set Maps Nonlinear Programming Comparative Study
Unable to display preview. Download preview PDF.
- J.W. Daniel, “Convergent step sizes for gradient-like feasible direction algorithms for constrained optimization,” in J.B. Rosen, O.L. Mangasarian and K. Ritter, eds., Nonlinear programming (Academic Press, New York, 1970) pp. 245–274.Google Scholar
- R.M. Elkin, “Convergence theorems for Gauss-Seidel and other minimization algorithms”, Computer Science Center Tech. Rept. 68-59, University of Maryland, College Park, MD (1968).Google Scholar
- G.G.L. Meyer, “Convergence conditions for a type of algorithm model”, Tech. Rept. 75-14, The Johns Hopkins University, Baltimore, MD (1975).Google Scholar
- R.R. Meyer, “A convergence theory for a class of anti-jamming strategies”, Tech. Rept. 1481, Math. Research Center, University of Wisconsin, Madison, WI. (1974).Google Scholar
- [21a]E. Polak, Computational methods in optimization: a unified approach (Academic Press, New York, 1971).Google Scholar
- E. Polak, “On the implementation of conceptual algorithms”, in J.B. Rosen, O.L. Mangasarian and K. Ritter, eds., Nonlinear programming (Academic Press, New York, 1970) pp. 275–291.Google Scholar
- S. Tishyadhigama, “General convergence theorems: Their relationships and applications”, Ph.D. Thesis, Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, CA, (1977).Google Scholar