Abstract
This paper is devoted to the study of gH-Clarke derivative for interval-valued functions. To find properties of the gH-Clarke derivative, the concepts of limit superior, limit inferior, and sublinear interval-valued functions are studied in the sequel. It is proved that the upper gH-Clarke derivative of a gH-Lipschitz continuous interval-valued function (IVF) always exists. For a convex and gH-Lipschitz IVF, the upper gH-Clarke derivative is found to be identical with the gH-directional derivative. It is observed that the upper gH-Clarke derivative is a sublinear IVF. Several numerical examples are provided to support the entire study.
Similar content being viewed by others
Availability of data and material
Not applicable.
References
Ansari QH, Lalitha CS, Mehta M (2013) Generalized convexity, nonsmooth variational inequalities, and nonsmooth optimization. CRC Press, New York
Bede B, Gal SG (2005) Generalizations of the differentiability of fuzzy-number-valued functions with applications to fuzzy differential equations. Fuzzy Sets Syst 151:581–599
Bhurjee AK, Padhan SK (2016) Optimality conditions and duality results for non-differentiable interval optimization problems. J Appl Math Comput 50(1–2):59–71
Chalco-Cano Y, Rufian-Lizana A, Román-Flores H, Jiménez-Gamero MD (2013) Calculus for interval-valued functions using generalized Hukuhara derivative and applications. Fuzzy Sets Syst 219:49–67
Chalco-Cano Y, Román-Flores H, Jiménez-Gamero MD (2011) Generalized derivative and \(\pi \)-derivative for set-valued functions. Inf Sci 181(11):2177–2188
Clarke FH (1990) Optimization and nonsmooth analysis, vol 5. SIAM
Costa TM, Chalco-Cano Y, Lodwick WA, Silva GN (2015) Generalized interval vector spaces and interval optimization. Inf Sci 311:74–85
Delfour MC (2012) Introduction to optimization and semidifferential calculus, Society for Industrial and Applied Mathematics
Demyanov VF (2002) The rise of nonsmooth analysis: its main tools. Cybern Syst Anal 38(4):527–547
Dutta J (2005) Generalized derivatives and nonsmooth optimization, a finite dimensional tour. TOP 13(2):185–279
Ghosh D (2017) Newton method to obtain efficient solutions of the optimization problems with interval-valued objective functions. J Appl Math Comput 53:709–731
Ghosh D, Ghosh D, Bhuiya SK, Patra LK (2018) A saddle point characterization of efficient solutions for interval optimization problems. J Appl Math Comput 58(1–2):193–217
Ghosh D (2016) A Newton method for capturing efficient solutions of interval optimization problems. Opsearch 53(3):648–665
Ghosh D, Chakraborty D (2019) An introduction to analytical fuzzy plane geometry, studies in fuzziness and soft computing, vol 381. Springer
Ghosh D, Chauhan RS, Mesiar R, Debnath AK (2020) Generalized Hukuhara Gâteaux and Fréchet derivatives of interval-valued functions and their application in optimization with interval-valued functions. Inf Sci 510:317–340
Ghosh D, Debnath AK, Pedrycz W (2020) A variable and a fixed ordering of intervals and their application in optimization with interval-valued functions. Int J Approx Reason 121:187–205
Ghosh D, Debnath AK, Chauhan RS, Castillo O (2020) Generalized-Hukuhara-Gradient efficient-direction method to solve optimization problems with interval-valued functions and its application in least squares problems. arXiv preprint arXiv:2011.10462
Guo Y, Ye G, Zhao D, Liu W (2019) \(gH\)-Symmetrically derivative of interval-Valued functions and applications in interval-valued optimization. Symmetry 11(10):1203
Hiriart-Urruty JB, Lemaréchal C (2012) Fundamentals of convex analysis. Springer Science & Business Media
Hukuhara M (1967) Intégration des applications measurables dont la valeur est un compact convexe. Funkcialaj Ekvacioj 10:205–223
Ishibuchi H, Tanaka H (1990) Multiobjective programming in optimization of the interval objective function. Eur J Oper Res 48(2):219–225
Jahn J (2007) Introduction to the theory of nonlinear optimization, 3rd edn. Springer Science and Business Media, New York
Kalani H, Akbarzadeh-T MR, Akbarzadeh A, Kardan I (2016) Interval-valued fuzzy derivatives and solution to interval-valued fuzzy differential equations. J Intell Fuzzy Syst 30(6):3373–3384
Lupulescu V (2013) Hukuhara differentiability of interval-valued functions and interval differential equations on time scales. Inf Sci 248:50–67
Lupulescu V (2015) Fractional calculus for interval-valued functions. Fuzzy Sets Syst 265:63–85
Markov S (1979) Calculus for interval functions of a real variable. Computing 22(4):325–337
Moore RE (1966) Interval analysis. Prentice-Hall, Englewood Cliffs
Moore RE (1987) Method and applications of interval analysis, Society for Industrial and Applied Mathematics
Ramík J, Vlach M (2002) Generalized concavity in optimization and decision making, vol 305. Kluwer Academic Publishers, Boston
Schirotzek W (2007) Nonsmooth analysis. Universitex, Springer Science & Business Media
Sengupta A, Pal TK, Chakraborty D (2001) Interpretation of inequality constraints involving interval coefficients and a solution to interval linear programming. Fuzzy Sets Syst 119(1):129–138
Stefanini L (2008) A generalization of Hukuhara difference. In Soft methods for handling variability and imprecision, advances in soft computing, pp 203–210
Stefanini L, Bede B (2009) Generalized Hukuhara differentiability of interval-valued functions and interval differential equations. Nonlinear Anal 71:1311–1328
Stefanini L, Bede B (2014) Generalized fuzzy differentiability with LU-parametric representation. Fuzzy Sets Syst 257:184–203
Stefanini L, Arana-Jiménez M (2019) Karush–Kuhn–Tucker conditions for interval and fuzzy optimization in several variables under total and directional generalized differentiability. Fuzzy Sets Syst 362:1–34
Van Hoa N (2015) The initial value problem for interval-valued second-order differential equations under generalized \(H\)-differentiability. Inf Sci 311:119–148
Wu HC (2007) The Karush–Kuhn–Tucker optimality conditions in an optimization problem with interval-valued objective function. Eur J Oper Res 176:46–59
Wu HC (2008) On interval-valued non-linear programming problems. J Math Anal Appl 338(1):299–316
Acknowledgements
We extend sincere thanks to the reviewers and editors for their valuable comments to improve the article. The first author is thankful for a research scholarship awarded by the University Grants Commission, Government of India.
Funding
Not applicable.
Author information
Authors and Affiliations
Contributions
All authors contributed to the study conception and analysis. Material preparation and analysis were performed by Ram Surat Chauhan, Debdas Ghosh, Jaroslav Ramík, and Amit Kumar Debnath. The first draft of the manuscript was written by Ram Surat Chauhan, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Code availability
Not applicable.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A Proof of Lemma 1
Proof
Let \(\mathbf{Z} =[\underline{z},~\overline{z}]\) and \(a, b \in {\mathbb {R}}\).
-
(i)
If \(\mathbf{Z} \succeq \mathbf{0} \), then
$$\begin{aligned} \underline{z}&\ge 0~\text {and}~\overline{z}\ge 0\\&\quad \implies |a |\underline{z}+|b |\underline{z} \ge |a+b |\underline{z}~\text {and}~|a |\overline{z}+|b |\overline{z} \ge |a+b |\overline{z}\\&\quad \implies |a+b |\odot \mathbf{Z} \preceq |a |\odot \mathbf{Z} \oplus |b |\odot \mathbf{Z} . \end{aligned}$$ -
(ii)
If \(\mathbf{Z} \preceq \mathbf{0} \), then
$$\begin{aligned}&\underline{z}\le 0~\text {and}~\overline{z}\le 0\\&\quad \implies |a |\underline{z}+|b |\underline{z} \le |a+b |\underline{z}~\text {and}~|a |\overline{z}+|b |\overline{z} \le |a+b |\overline{z}\\&\quad \implies |a+b |\odot \mathbf{Z} \succeq |a |\odot \mathbf{Z} \oplus |b |\odot \mathbf{Z} . \end{aligned}$$ -
(iii)
If \(\mathbf{Z} \nprec \mathbf{0} \), then
$$\begin{aligned} \overline{z}&\ge 0 \implies |a |\overline{z}+|b |\overline{z} \ge |a+b |\overline{z}\\&\quad \implies |a+b |\odot \mathbf{Z} \nsucc |a |\odot \mathbf{Z} \oplus |b |\odot \mathbf{Z} . \end{aligned}$$
\(\square \)
Appendix B Proof of Lemma 2
Proof
Let \(\mathbf{A} = [\underline{a}, \overline{a}], ~\mathbf{B} = [\underline{b}, \overline{b}],~\mathbf{C} = [\underline{c}, \overline{c}]\) and \(\mathbf{D} = [\underline{d}, \overline{d}]\).
-
(i)
We have the following four possible cases.
-
Case 1. Let \(\overline{a}-\overline{c}\ge \underline{a}-\underline{c}\) and \(\overline{c}-\overline{b}\ge \underline{c}-\underline{b}\). Then, \(\overline{a}-\overline{b}\ge \underline{a}-\underline{b}\) and
$$\begin{aligned} (\mathbf{A} \ominus _{gH}{} \mathbf{C} )\oplus (\mathbf{C} \ominus _{gH}{} \mathbf{B} )= & {} [\underline{a}-\underline{c}, \overline{a}-\overline{c}]\oplus [\underline{c}-\underline{b},\overline{c}-\overline{b}]\\= & {} [\underline{a}-\underline{b},\overline{a}-\overline{b}]=\mathbf{A} \ominus _{gH}{} \mathbf{B} . \end{aligned}$$ -
Case 2. Let \(\overline{a}-\overline{c}\le \underline{a}-\underline{c}\) and \(\overline{c}-\overline{b}\le \underline{c}-\underline{b}\). Therefore, \(\overline{a}-\overline{b}\le \underline{a}-\underline{b}\) and
$$\begin{aligned}&(\mathbf{A} \ominus _{gH}{} \mathbf{C} )\oplus (\mathbf{C} \ominus _{gH}{} \mathbf{B} )\\&\quad = [\overline{a}-\overline{c}, \underline{a}-\underline{c}]\oplus [\overline{c}-\overline{b},\underline{c}-\underline{b}]\\&\quad =[\overline{a}-\overline{b},\underline{a}-\underline{b}]=\mathbf{A} \ominus _{gH}{} \mathbf{B} . \end{aligned}$$ -
Case 3. Let \(\overline{a}-\overline{c}<\underline{a}-\underline{c}\) and \(\overline{c}-\overline{b}>\underline{c}-\underline{b}\). Therefore,
$$\begin{aligned}&(\mathbf{A} \ominus _{gH}{} \mathbf{C} )\oplus (\mathbf{C} \ominus _{gH}{} \mathbf{B} )\\&\quad = [\overline{a}-\overline{c}, \underline{a}-\underline{c}]\oplus [\underline{c}-\underline{b},\overline{c}-\overline{b}]\\&\quad =[\overline{a}-\overline{c}+\underline{c}-\underline{b}, \underline{a}-\underline{c}+\overline{c}-\overline{b}]. \end{aligned}$$If possible, let
$$\begin{aligned} (\mathbf{A} \ominus _{gH}{} \mathbf{C} )\oplus (\mathbf{C} \ominus _{gH}{} \mathbf{B} ) \prec \mathbf{A} \ominus _{gH}{} \mathbf{B} .\nonumber \\ \end{aligned}$$(Appendix B.1)If \(\overline{a}-\overline{b}\ge \underline{a}-\underline{b}\), then from (Appendix B.1) we get
$$\begin{aligned}&[\overline{a}-\overline{c}+\underline{c}-\underline{b}, \underline{a}-\underline{c}+\overline{c}-\overline{b}] \prec [\underline{a}-\underline{b},\overline{a}-\overline{b}]\\&\quad \Longrightarrow \underline{a}-\underline{c}+\overline{c}-\overline{b} \le \overline{a}-\overline{b}\\&\quad \Longrightarrow \underline{a}-\underline{c} \le \overline{a}-\overline{c},~\text { which is an impossibility}. \end{aligned}$$Further, if \(\overline{a}-\overline{b}\le \underline{a}-\underline{b}\), then from (Appendix B.1), we have
$$\begin{aligned}&[\overline{a}-\overline{c}+\underline{c}-\underline{b}, \underline{a}-\underline{c}+\overline{c}-\overline{b}] \prec [\overline{a}-\overline{b},\underline{a}-\underline{b}]\\&\quad \Longrightarrow \underline{a}-\underline{c}+\overline{c}-\overline{b} \le \underline{a}-\underline{b}\\&\quad \Longrightarrow \overline{c}-\overline{b} \le \underline{c}-\underline{b},~\text {which is an impossibility}. \end{aligned}$$Thus, (Appendix B.1) is not true.
-
Case 4. Let \(\overline{a}-\overline{c}>\underline{a}-\underline{c}\) and \(\overline{c}-\overline{b}<\underline{c}-\underline{b}\). Proceeding as in Case 3 of (i) we can prove that (Appendix B.1) is not true. Hence,
$$\begin{aligned} (\mathbf{A} \ominus _{gH}{} \mathbf{C} )\oplus (\mathbf{C} \ominus _{gH}{} \mathbf{B} ) \nprec \mathbf{A} \ominus _{gH}{} \mathbf{B} . \end{aligned}$$ -
-
(ii)
As \({\Vert \mathbf{B} \ominus _{gH} \mathbf{A} \Vert }_{I({\mathbb {R}})} = \max \{|\underline{b}-\underline{a}|, |\overline{b}-\overline{a}|\},\) we break the proof in two cases.
-
Case 1. If \((L = )~ {\Vert \mathbf{B} \ominus _{gH} \mathbf{A} \Vert }_{I({\mathbb {R}})} = |\underline{b}-\underline{a}|\), then
$$\begin{aligned} |\underline{b}-\underline{a}| \ge |\overline{b}-\overline{a}|&\implies |\underline{b}-\underline{a}| \ge \overline{b}-\overline{a} \nonumber \\&\implies \overline{b} \le \overline{a}+L.\nonumber \\ \end{aligned}$$(Appendix B.2)Since \( \underline{b}-\underline{a} \le |\underline{b}-\underline{a}|\), then
$$\begin{aligned} \underline{b} \le \underline{a}+L. \end{aligned}$$(Appendix B.3)From (Appendix B.2) and (Appendix B.3), we have \(\mathbf{B} \preceq \mathbf{A} \oplus [L, L].\)
-
Case 2. If \((L = )~ {\Vert \mathbf{B} \ominus _{gH} \mathbf{A} \Vert }_{I({\mathbb {R}})} = |\overline{b}-\overline{a}|\), then
$$\begin{aligned}&|\underline{b}-\underline{a}| \le |\overline{b}-\overline{a}|\nonumber \\&\quad \implies \underline{b}-\underline{a} \le |\overline{b}-\overline{a}| \implies \underline{b} \le \underline{a}+L. \nonumber \\ \end{aligned}$$(Appendix B.4)Since \( \overline{b}-\overline{a} \le |\overline{b}-\overline{a}|\),
$$\begin{aligned} \overline{b} \le \overline{a}+L. \end{aligned}$$(Appendix B.5)From (Appendix B.4) and (Appendix B.5), we obtain \(\mathbf{B} \preceq \mathbf{A} \oplus [L, L], ~\text {where}~ L=\Vert \mathbf{B} \ominus _{gH}{} \mathbf{A} \Vert _{I({\mathbb {R}})}.\)
-
-
(iii)
If possible, let there exist \(\mathbf{A} ,~\mathbf{B} ,~\mathbf{C} \) and \(\mathbf{D} \) in \(I({\mathbb {R}})\) such that
$$\begin{aligned}&{\Vert (\mathbf{A} \ominus _{gH}{} \mathbf{B} )\ominus _{gH} (\mathbf{C} \ominus _{gH}{} \mathbf{D} ) \Vert }_{I({\mathbb {R}})}\nonumber \\&\quad > \Vert \mathbf{A} \ominus _{gH}{} \mathbf{C} \Vert _{I({\mathbb {R}})} + \Vert \mathbf{B} \ominus _{gH}{} \mathbf{D} \Vert _{I({\mathbb {R}})}.\nonumber \\ \end{aligned}$$(Appendix B.6)According to the definition of gH-difference of two intervals,
$$\begin{aligned}&\text {either}~~ \mathbf{A} \ominus _{gH} \mathbf{B} = [\underline{a}-\underline{b}, \overline{a}-\overline{b}]\\&\quad ~~\text {or}~~ \mathbf{A} \ominus _{gH} \mathbf{B} = [\overline{a}-\overline{b}, \underline{a}-\underline{b}], \\&\quad \text {either}~~ \mathbf{C} \ominus _{gH} \mathbf{D} = [\underline{c}-\underline{d}, \overline{c}-\overline{d}]~~\\&\quad \text {or}~~\mathbf{C} \ominus _{gH} \mathbf{D} = [\overline{c}-\overline{d}, \underline{c}-\underline{d}], \\&\quad \text {either}~~ \mathbf{A} \ominus _{gH} \mathbf{C} = [\underline{a}-\underline{c}, \overline{a}-\overline{c}]\\&\quad \text {or}~~ \mathbf{A} \ominus _{gH} \mathbf{C} = [\overline{a}-\overline{c}, \underline{a}-\underline{c}], \end{aligned}$$and
$$\begin{aligned}&\text {either}~~ \mathbf{B} \ominus _{gH} \mathbf{D} = [\underline{b}-\underline{d}, \overline{b}-\overline{d}]~~\text {or}~~\mathbf{B} \ominus _{gH} \mathbf{D} \\&\quad = [\overline{b}-\overline{d}, \underline{b}-\underline{d}]. \end{aligned}$$Then, one of the following holds true:
-
(a)
\((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\underline{a}-\underline{b}-\underline{c}+\underline{d},~ \overline{a}-\overline{b}-\overline{c}+\overline{d} ]\)
-
(b)
\((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\underline{a}-\underline{b}-\overline{c}+\overline{d},~ \overline{a}-\overline{b}-\underline{c}+\underline{d} ]\)
-
(c)
\((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [ \overline{a}-\overline{b}-\overline{c}+\overline{d},~\underline{a}-\underline{b}-\underline{c}+\underline{d} ]\)
-
(d)
\((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\overline{a}-\overline{b}-\underline{c}+\underline{d},~ \underline{a}-\underline{b}-\overline{c}+\overline{d} ]\).
-
Case 1. Let \((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\underline{a}-\underline{b}-\underline{c}+\underline{d},~ \overline{a}-\overline{b}-\overline{c}+\overline{d} ]\).
-
(a)
If \({\Vert (\mathbf{A} \ominus _{gH}{} \mathbf{B} )\ominus _{gH} (\mathbf{C} \ominus _{gH}{} \mathbf{D} ) \Vert }_{I({\mathbb {R}})} = |\underline{a}-\underline{b}-\underline{c}+\underline{d} |\), then from equation (Appendix B.6), we have
$$\begin{aligned} |\underline{a}-\underline{b}-\underline{c}+\underline{d} |> & {} |\underline{a}-\underline{c}|+ |\underline{b}-\underline{d}|\\> & {} |\underline{a}-\underline{b}-\underline{c}+\underline{d} |, \end{aligned}$$which is impossible.
-
(b)
If \({\Vert (\mathbf{A} \ominus _{gH}{} \mathbf{B} )\ominus _{gH} (\mathbf{C} \ominus _{gH}{} \mathbf{D} ) \Vert }_{I({\mathbb {R}})} = |\overline{a}-\overline{b}-\overline{c}+\overline{d} |\), then from equation (Appendix B.6), we have
$$\begin{aligned} |\overline{a}-\overline{b}-\overline{c}+\overline{d} |> & {} |\overline{a}-\overline{c}|+ |\overline{b}-\overline{d}|\\> & {} |\overline{a}-\overline{b}-\overline{c}+\overline{d} |, \end{aligned}$$which is again impossible.
-
(a)
-
Case 2. Let \((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\overline{a}-\overline{b}-\overline{c}+\overline{d},~ \underline{a}-\underline{b}-\underline{c}+\underline{d} ]\). For this case, two subcases are similar to the Case 1 of (iii) will lead to impossibilities.
-
Case 3. Let \((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\underline{a}-\underline{b}-\overline{c}+\overline{d},~ \overline{a}-\overline{b}-\underline{c}+\underline{d} ]\). Then,
$$\begin{aligned} \underline{a}-\underline{b} \le \overline{a}-\overline{b}~\text {and}~\overline{c}-\overline{d} \le \underline{c}-\underline{d}.\nonumber \\ \end{aligned}$$(Appendix B.7)-
(a)
If \({\Vert (\mathbf{A} \ominus _{gH}{} \mathbf{B} )\ominus _{gH} (\mathbf{C} \ominus _{gH}{} \mathbf{D} ) \Vert }_{I({\mathbb {R}})} = |\overline{a}-\overline{b}-\underline{c}+\underline{d} |\), then \(\overline{a}-\overline{b}-\underline{c}+\underline{d} \ge 0.\) From equation (Appendix B.6), we have
$$\begin{aligned}&|\overline{a}-\overline{b}-\underline{c}+\underline{d} |> |\overline{a}-\overline{c}|+ |\overline{b}-\overline{d}|\\&\quad \implies \overline{c}-\overline{d} > \underline{c}-\underline{d}, \end{aligned}$$which is contradictory to (Appendix B.7).
-
(b)
If \({\Vert (\mathbf{A} \ominus _{gH}{} \mathbf{B} )\ominus _{gH} (\mathbf{C} \ominus _{gH}{} \mathbf{D} ) \Vert }_{I({\mathbb {R}})} = |\underline{a}-\underline{b}-\overline{c}+\overline{d} |\), then \(\underline{a}-\underline{b}-\overline{c}+\overline{d} < 0.\) From equation (Appendix B.6), we have
$$\begin{aligned}&-(\underline{a}-\underline{b}-\overline{c}+\overline{d}) = |\underline{a}-\underline{b}-\overline{c}+\overline{d} |> |\underline{a}\\&\quad -\underline{c}|+ |\underline{b}-\underline{d}|\implies \overline{c}-\overline{d} > \underline{c}-\underline{d}, \end{aligned}$$which is again contradictory to (Appendix B.7).
-
(a)
-
Case 4. Let \((\mathbf{A} \ominus _{gH} \mathbf{B} )\ominus _{gH}(\mathbf{C} \ominus _{gH} \mathbf{D} )= [\overline{a}-\overline{b}-\underline{c}+\underline{d},~\underline{a}-\underline{b}-\overline{c}+\overline{d} ]\). All the two subcases for this case are similar to Case 3 of (iii).
Hence, (Appendix B.6) is wrong, and thus the result follows.
-
(a)
\(\square \)
Rights and permissions
About this article
Cite this article
Chauhan, R.S., Ghosh, D., Ramík, J. et al. Generalized Hukuhara-Clarke Derivative of Interval-valued Functions and its Properties. Soft Comput 25, 14629–14643 (2021). https://doi.org/10.1007/s00500-021-06251-w
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00500-021-06251-w