×

Well-posedness criteria in optimization with application to the calculus of variations. (English) Zbl 0841.49005

In this paper, for the global optimization problem \((X, J)\), to minimize the proper extended real-valued function \(J: X\to (- \infty, \infty)\) over the given subset \(X\) of a normed linear space equipped with the strong convergence, well-posedness criteria are derived. The given problem is embedded into a smoothly parametrized family \((X, I(., p))\) of minimization problems, where \(p\) is a parameter belonging to a given Banach space \(L\), and \(p^*\) is the parameter value to which the given unperturbed problem corresponds, i.e., \(I(x, p^*)= J(x)\) \(\forall x\in X\). Defining the value function \(V(p)= \inf\{I(x, p)\mid x\in X\}\) the author gives the following definition of well-posedness.
\((X, J)\) is well-posed with respect to the embedding iff \(V(p)> -\infty\), \(\forall p\in L\), and there exists a unique \(x^*= \arg\min(X, J)\) and for every sequence \(p_n\to p^*\) and every sequence \(x_n\in X\) such that \(I(x_n, p_n)- V(p_n)\to 0\) as \(n\to \infty\) we have \(x_n\to x^*\) in \(X\).
This definition is stronger than the Tikhonov well-posedness. In the following, the defined well-posedness is related under suitable conditions to the differentiability properties of \(V\) at \(p^*\). These abstract results are applied to one-dimensional problems of the calculus of variations.

MSC:

49J27 Existence theories for problems in abstract spaces
49K99 Optimality conditions
90C99 Mathematical programming
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Asplund, E.; Rockafellar, R. T., Gradients of convex functions, Trans. Am. math. Soc., 139, 443-467 (1969) · Zbl 0181.41901
[2] Fitzpatrick, S., Metric projections and the differentiability of distance functions, Bull. Aust. math. Soc., 22, 291-312 (1980) · Zbl 0437.46012
[3] Fleming, W. H., The Cauchy problem for a nonlinear first order partial differential equation, J. diff. Eqns, 5, 515-530 (1969) · Zbl 0172.13901
[4] Fleming, W. H.; Rishel, R., (Deterministic and Stochastic Optimal Control (1975), Springer)
[5] Clarke, F. H.; Loewen, P. D., The value function in optimal control: sensitivity, controllability and time-optimality, SIAM J. Control Optim., 24, 243-263 (1986) · Zbl 0601.49020
[6] Clarke, F. H., (Optimization and Nonsmooth Analysis (1983), Wiley: Wiley Berlin) · Zbl 0582.49001
[7] Ekeland, I.; Lebourg, G., Generic Fréchet differentiability and perturbed optimization problems in Banach spaces, Trans. Am. math. Soc., 224, 193-216 (1976) · Zbl 0313.46017
[8] Georgiev, P. G., Well posed optimization problems depending on a parameter and differentiability of the optimal value function, (Extended abstracts of the Workshop on well-posedness in optimization. Extended abstracts of the Workshop on well-posedness in optimization, S. Margherita Ligure, 19-21 (1991), Dipartimento di Matematica, Univ. di Genova: Dipartimento di Matematica, Univ. di Genova New York)
[9] Henry, D., Nonuniqueness of solutions in the calculus of variations: a geometric approach, SIAM J. Control. Optim., 18, 627-639 (1980) · Zbl 0454.49008
[10] Fleming, W. H.; Soner, H. M., (Controlled Markov Processes and Viscosity Solutions (1993), Springer) · Zbl 0773.60070
[11] Dontchev, A. L.; Zolezzi, T., (Well-posed Optimization Problems, Lecture Notes in Mathematics, Vol. 1543 (1993), Springer: Springer Berlin) · Zbl 0797.49001
[12] Kuratowski, C., (Topologie, Vol. 1 (1958), Panstwowe Wydawnictwo Naukowe: Panstwowe Wydawnictwo Naukowe New York)
[13] Ekeland, I., Nonconvex minimization problems, Bull. Am. math. Soc., 1, 443-474 (1979) · Zbl 0441.49011
[14] Kutznetzov, N. N.; Siskin, A. A., On a many dimensional problem in the theory of quasilinear equations, Zh. vychisl. mat. Fiz., 4, 192-205 (1964)
[15] Cesari, L., (Optimization-Theory and Applications (1983), Springer: Springer Warsaw)
[16] Marcellini, P., (Nonconvex Integrals of the Calculus of Variations, Lecture Notes in Mathematics, Vol. 1446 (1990), Springer: Springer Berlin), 16-57 · Zbl 0735.49002
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.