×

Existence and iterative approximations of solutions for certain functional equation and inequality. (English) Zbl 1293.49057

The paper studies a functional equation and inequality which arise in dynamic programming of multi-stage decision processes. The authors are interested in the solvability of and iterative approximations for the solutions to the functional equation and inequality.
Let \(S\subset X\) and \(D\subset Y\) be the state and decision spaces, respectively, where \(X\) and \(Y\) are real Banach spaces. The authors study the functional equation and inequality stated below, where \(opt\) denotes \(sup\) or \(inf\), \(\lambda \) is a constant in \([0,1],\) \(x\) and \(y\), respectively, represent the state and decision vectors, \(f\left( x\right) \) denotes the optimal return function with initial state \(x\), and \(a\) and \(b\) denote transformation processes. \[ \begin{split} f\left( x\right) =\lambda \text{ }opt_{y\in D}\{u(x,y)+A(x,y,f(a(x,y))\}\\ +(1-\lambda )opt_{y\in D}\{v(x,y)+B(x,y,f(b(x,y))\},\forall x\in S \end{split}\tag{1} \]
\[ \begin{split} f\left( x\right) \geq \lambda \text{ }opt_{y\in D}\{u(x,y)+A(x,y,f(a(x,y))\}\\ +(1-\lambda )opt_{y\in D}\{v(x,y)+B(x,y,f(b(x,y))\},\forall x\in S \end{split}\tag{2} \] Let \(F\) denote the set of functions \(f\) from \(S\) to \(\mathbb{R}\equiv ]-\infty ,+\infty [ .\) Define the sets \(B\left( S\right),BC\left( S\right) ,BB\left( S\right) \) as follows: \[ \begin{aligned} & B\left( S\right) \equiv \left\{ f\in F:f\text{ }is\text{ }bounded\right\}\\ & BC\left( S\right) \equiv \{f\in B\left( S\right) :f\text{ }is\text{ } continuous\} \\ & BB\left( S\right) \equiv \{f\in F:f\text{ }is\text{ }bounded\text{ }on \text{ }each\text{ }bounded\text{ }subset\text{ }of\text{ }S\}. \end{aligned} \] Define the norm \(\|\cdot \|_{1}\)on \(B\left( S\right) \) and \(BC\left( S\right) \) by \(\|w\|_{1}\equiv \sup_{x\in S}|w\left( x\right)|\). Then \(\left( B\left( S\right) ,\|\cdot\|_{1}\right) \) and \(\left( BC\left( S\right) ,\|\cdot\|_{1}\right) \) are Banach spaces.
Also, denoting by \(\mathbb{N}\) the set of positive integers, define, \(\forall \left( k,f,g\right) \in\mathbb{N}\times BB\left( S\right) \times BB\left(S\right)\), \[ \begin{aligned} &\overline{B}\left( 0,k\right) \equiv \left\{ x\in S:\|x\|\leq k\right\} \\ & d_{k}\left( f,g\right) \equiv \sup \left\{ \mid f\left( x\right) -g\left(x\right) \mid :x\in \overline{B}\left( 0,k\right) \right\} \\ & d\left( f,g\right) \equiv \sum_{k=1}^{\infty }\frac{1}{2^{k}}\cdot \frac{d_{k}\left( f,g\right) }{1+d_{k}\left( f,g\right) } \end{aligned} \] Some preliminary lemmas are proved to establish that \(\left( BB\left(S\right) ,d\right) \) is a complete metric space induced by the countable family of pseudo-metrics \(\{d_{k}\}_{k\in\mathbb{N}}\).
Using several fixed-point theorems due to Krasnoselskii, Boyd–Wong and Liu, the paper goes on to prove theorems establishing the existence and/or uniqueness and iterative approximations of solutions for the functional equation \((1)\) in the Banach spaces \(BC(S)\) and \(B(S)\) and the complete metric space \(BB(S)\), respectively. Examples are provided to illustrate the usefulness of the theorems.
In the final section, utilizing the monotone iterative method, it goes on to prove a couple of theorems regarding the solvability and iterative approximations of the functional inequality \((2)\) in the complete metric space \((BB(S),d),\) and provide some illustrative examples.
The functional equation studied in the paper includes the functional equations in some of the earlier literature as special cases. The results of the paper extend, improve and unify some earlier results in the literature.

MSC:

49L20 Dynamic programming in optimal control and differential games
49J27 Existence theories for problems in abstract spaces
39B22 Functional equations for real functions
49L99 Hamilton-Jacobi theories
90C39 Dynamic programming
90C48 Programming in abstract spaces
47N10 Applications of operator theory in optimization, convex analysis, mathematical programming, economics
Full Text: DOI

References:

[1] Bellman, R.: Some functional equations in the theory of dynamic programming, I. Functions of points and point transformations. Trans. Am. Math. Soc. 80, 55-71 (1955) · Zbl 0066.13704 · doi:10.1090/S0002-9947-1955-0074692-0
[2] Bellman, R.: Dynamic Programming. Princeton University Press, Princeton (1957) · Zbl 0077.13605
[3] Bellman, R., Lee, E.S.: Functional equations arising in dynamic programming. Aequ. Math. 17, 1-18 (1978) · Zbl 0397.39016 · doi:10.1007/BF01818535
[4] Bellman, R., Roosta, M.: A technique for the reduction of dimensionality in dynamic programming. J. Math. Anal. Appl. 88, 543-546 (1982) · Zbl 0498.90079 · doi:10.1016/0022-247X(82)90212-8
[5] Bhakta, P.C., Choudhury, S.R.: Some existence theorems for functional equations arising in dynamic programming II. J. Math. Anal. Appl. 131, 217-231 (1988) · Zbl 0662.90085 · doi:10.1016/0022-247X(88)90201-6
[6] Bhakta, P.C., Mitra, S.: Some existence theorems for functional equations arising in dynamic programming. J. Math. Anal. Appl. 98, 348-362 (1984) · Zbl 0533.90091 · doi:10.1016/0022-247X(84)90254-3
[7] Liu, Z.: Coincidence theorems for expansion mappings with applications to the solutions of functional equations arising in dynamic programming. Acta Sci. Math. 65, 359-369 (1999) · Zbl 0929.54036
[8] Liu, Z.: Compatible mappings and fixed points. Acta Sci. Math. 65, 371-383 (1999) · Zbl 0928.54040
[9] Liu, Z.: Existence theorems of solutions for certain classes of functional equations arising in dynamic programming. J. Math. Anal. Appl. 262, 529-553 (2001) · Zbl 1018.90064 · doi:10.1006/jmaa.2001.7551
[10] Liu, Z., Agarwal, R.P., Kang, S.M.: On solvability of functional equations and system of functional equations arising in dynamic programming. J. Math. Anal. Appl. 297, 111-130 (2004) · Zbl 1055.39038 · doi:10.1016/j.jmaa.2004.04.049
[11] Liu, Z., Kang, S.M.: Properties of solutions for certain functional equations arising in dynamic programming. J. Glob. Optim. 34, 273-292 (2006) · Zbl 1091.49021 · doi:10.1007/s10898-005-2605-6
[12] Liu, Z., Kang, S.M.: Existence and uniqueness of solutions for two classes of functional equations arising in dynamic programming. Acta Math. Appl. Sin. 23, 195-208 (2007) · Zbl 1175.49024 · doi:10.1007/s10255-007-0363-6
[13] Liu, Z., Kang, S.M., Ume, J.S.: Solvability and convergence of iterative algorithms for certain functional equations arising in dynamic programming. Optimization 59(6), 887-916 (2010) · Zbl 1198.49023 · doi:10.1080/02331930902884182
[14] Liu, Z., Ume, J.S.: On properties of solutions for a class of functional equations arising in dynamic programming. J. Optim. Theory Appl. 117, 533-551 (2003) · Zbl 1045.90075 · doi:10.1023/A:1023945621360
[15] Liu, Z., Ume, J.S., Kang, S.M.: Some existence theorems for functional equations arising in dynamic programming. J. Korean Math. Soc. 43, 11-28 (2006) · Zbl 1091.49022 · doi:10.4134/JKMS.2006.43.1.011
[16] Liu, Z., Ume, J.S., Kang, S.M.: Some existence theorems for functional equations and system of functional equations arising in dynamic programming. Taiwan. J. Math. 14(4), 1517-1536 (2010) · Zbl 1215.49008
[17] Liu, Z., Xu, Y.G., Ume, J.S., Kang, S.M.: Solutions to two functional equations arising in dynamic programming. J. Comput. Appl. Math. 192, 251-269 (2006) · Zbl 1149.65097 · doi:10.1016/j.cam.2005.04.033
[18] Liu, Z., Zhao, L.S., Kang, S.M., Ume, J.S.: On the solvability of a functional equation. Optimization 60(3), 365-375 (2011) · Zbl 1217.49026 · doi:10.1080/02331930903121311
[19] Boyd, D.W., Wong, J.S.W.: On nonlinear contractions. Proc. Am. Math. Soc. 20, 458-464 (1969) · Zbl 0175.44903 · doi:10.1090/S0002-9939-1969-0239559-9
[20] Erbe, L.H., Kong, Q.K., Zhang, B.G.: Oscillation Theory for Functional Differential Equations. Marcel Dekker, New York (1995) · Zbl 0821.34067
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.