Any weakly decreasing solution of the Hamilton–Jacobi inequality generates a so-called accessory problem of dynamic optimization over Krasovskii–Subbotin constructive motions (Euler curves) produced by extremal feedback control strategies. We derive conditions under which an optimal trajectory of the considered Mayer optimal control problem is a minimizer of the accessory problem for a fixed majorant — a certain solution of the Hamilton–Jacobi inequality. The result is formulated in terms of the compatibility of the latter solution with an optimal trajectory. In the general case of a nonsmooth majorant (and a nonsmooth problem), the optimality condition means that there is a component of the proximal subdifferential of the majorant along the optimal trajectory that coincides with a certain solution of an adjoint inclusion arising in the maximum principle of Kaskosz and Łojasiewicz. This implies the general feedback minimum principle — a global necessary optimality condition, which strengthens all known formulations of the maximum principle for problems without terminal constraints.
Keywords: extremals, feedback controls, weakly decreasing functions, feedback minimum principle
Received June 14, 2022
Revised June 30, 2022
Accepted July 4, 2022
Vladimir Aleksandrovich Dykhta, Dr. Phys.-Math. Sci., Prof., Matrosov Institute for System Dynamics and Control Theory of Siberian Branch of Russian Academy of Sciences, Irkutsk, 664033 Russia, e-mail: dykhta@gmail.com
REFERENCES
1. Clarke P.H., Ledyaev Yu.S., Stern R.J., Wolenski P.R. Qualitative properties of trajectories of control systems: A survey. J. Dyn. Control Syst., 1995, vol. 1, no. 1, pp. 1–48. doi: 10.1007/BF02254655
2. Clarke F.H., Ledyaev Yu.S., Stern R.J., Wolenski R.R. Nonsmooth analysis and control theory. NY: Springer, 1998, 278 p. doi: 10.1007/b97650 .
3. Krasovskii N.N., Subbotin A.I. Pozitsionnye differentsial’nye igry [Positional differential games]. Moscow: Nauka Publ., 1974, 458 p.
4. Subbotin A.I. Generalized solutions of first-order PDEs: The dynamical optimization perspective. Basel: Birkhäuser, 1995, 314 p. doi: 10.1007/978-1-4612-0847-1 . Translated to Russian under the title Obobshchennye resheniya uravnenii v chastnykh proizvodnykh pervogo poryadka: Perspektivy dinamicheskoi optimizatsii, Moscow, Izhevsk: Inst. Komp’yuter. Issled., 2003, 336 p.
5. Clarke F., Ledyaev Yu.S., Subbotin A.I. Universal feedback control via proximal aiming in problems of control under disturbance and differential games. Proc. Steklov Inst. Math., 1999, vol. 224, pp. 149–168.
6. Dykhta V.A. Weakly monotone solutions of the Hamilton–Jacobi inequality and optimality conditions with positional controls. Autom. Remote Control, 2014, vol. 75, no. 5, pp. 829–844. doi: 10.1134/S0005117914050038
7. Dykhta V.A. Nonstandard duality and nonlocal necessary optimality conditions in nonconvex optimal control problems. Autom. Remote Control, 2014, vol. 75, no. 11, pp. 1906–1921. doi: 10.1134/S0005117914110022
8. Dykhta V.A. Variational necessary optimality conditions with feedback descent controls for optimal control problems. Dokl. Math., 2015, vol. 91, no. 3, pp. 394–396. doi: 10.1134/S106456241503031X
9. Dykhta V.A. Approximate feedback minimum principle for suboptimal processes in non-smooth optimal control problems. In: Proc. Int. Conf. “Stability, Control and Differential Games” (SCDG2019, September 16–20, 2019). Cham: Springer, 2020, pp. 127–132. doi: 10.1007/978-3-030-42831-0_12
10. Kaśkosz B., Lojasiewicz S. A maximum principle for generalized control systems. Nonlinear Analysis: Theory Methods & Applications, 1985, vol. 9, no. 2, pp. 109–130. doi: 10.1016/0362-546X(85)90067-7
11. Kaśkosz B. Extremality, controllability, and abundant subsets of generalized control systems. J. Optim. Theory Appl., 1999, vol. 101, no. 1, pp. 73–108. doi: 10.1023/A:1021719027140
12. Frankowska H., Kaśkosz B. Linearization and boundary trajectories of nonsmooth control systems. Can. J. Math., 1988, vol. 40, no. 3, pp. 589–609. doi: 10.4153/CJM-1988-025-7
13. Sussmann H. A strong version of the Lojasiewicz maximum principle. In: Optimal Control of Differential Equations, ed. N.H. Pavel, Ser. Lecture Notes in Pure and Applied Mathematics, N.Y.: M. Dekker, 1994, pp. 1–17. ISBN: 9781003072225
14. Loewen P.D., Vinter R. Pontryagin-type necessary conditions for differential inclusion problems. Systems & Control Letters, 1987, vol. 9, no. 3, pp. 263–265. doi: 10.1016/0167-6911(87)90049-1
15. Artstein Z. Pontryagin maximum principle revisited with feedbacks. Eur. J. Control, 2011, vol. 17, no. 1, pp. 46–54. doi: 10.3166/ejc.17.46-54
16. Dykhta V.A., Samsonyuk O.N. Neravenstva Gamil’tona–Yakobi i variatsionnye usloviya optimal’nosti [Hamilton–Jacobi inequalities and variational optimality conditions]. Irkutsk: Irkutskii Gos. Universitet Publ., 2015, 150 p. ISBN: 978-5-9624-1298-6 .
17. Dykhta V.A. Positional strengthenings of the maximum principle and sufficient optimality conditions. Proc. Steklov Inst. Math., 2016, vol. 293, suppl. 1, pp. 43–57. doi: 10.1134/S0081543816050059
18. Clarke H. Optimization and nonsmooth analysis. Philadelphia: SIAM, 1987, 320 p. ISBN: 0898712564 . Translated to Russian under the title Optimizatsiya i negladkii analiz, Moscow: Nauka Publ., 1988, 280 p.
19. Aubin J.-P., Frankowska H. Set-valued analysis. Boston: Birkhäuser, 1990, 461 p. ISBN: 0817634789 .
20. Bardi M., Capuzzo-Dolcetta I. Optimal control and viscosity solutions of Hamilton–Jacobi–Bellman equations. Boston: Birkhäuser, Springer, 1997, 574 p. doi: 10.1007/978-0-8176-4755-1
Cite this article as: V.A. Dykhta. On the set of necessary optimality conditions with positional controls generated by weakly decreasing solutions of the Hamilton–Jacobi inequality. Trudy Instituta Matematiki i Mekhaniki UrO RAN, 2022, vol. 28, no. 3, pp. 83–93.