Abstract
We propose a new numerical method for semi-infinite optimization problems whose objective function is a nonsmooth function. Existing numerical methods for solving semi-infinite programming (SIP) problems make strong assumptions on the structure of the objective function, e.g., differentiability, or are not guaranteed to furnish a feasible point on finite termination. In this paper, we propose a feasible proximal bundle method with convexification for solving this class of problems. The main idea is to derive upper bounding problems for the SIP by replacing the infinite number of constraints with a finite number of convex relaxation constraints, introduce improvement functions for the upper bounding problems, construct cutting plane models of the improvement functions, and reformulate the cutting plane models as quadratic programming problems and solve them. The convex relaxation constraints are constructed with ideas from the α BB method of global optimization. Under mild conditions, we showed that every accumulation point of the iterative sequence is an 𝜖-stationary point of the original SIP problem. Under slightly stronger assumptions, every accumulation point of the iterative sequence is a local solution of the original SIP problem. Preliminary computational results on solving nonconvex, nonsmooth constrained optimization problems and semi-infinite optimization problems are reported to demonstrate the good performance of the new algorithms.


We’re sorry, something doesn't seem to be working properly.
Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.
References
Adjiman, C.S., Androulakis, I.P., Floudas, C.A.: A global optimization method, α BB, for general twice-differentiable constrained NLPs-I: theoretical advances. Comput. Chem. Eng. 22, 1137–1158 (1998)
Adjiman, C.S., Androulakis, I.P., Floudas, C.A.: A global optimization method, α BB, for general twice-differentiable constrained NLPs-II: implementation and computational results. Comput. Chem. Eng. 22, 1159–1179 (1998)
Bagirov, A., Karmitsa, N., Mäkelä, M. M.: Introduction to Nonsmooth Optimization: Theory, Practice and Software. Springer, Cham (2014)
Bhattacharjee, B., Green, Jr.W. H., Barton, P.I.: Interval methods for semi-infinite programs. Comput. Optim. Appl. 30, 63–93 (2005)
Bhattacharjee, B., Lemonidis, P., Green, W.H. Jr., Barton, P.I.: Global solution of semi-infinite programs. Math. Program. 103, 283–307 (2005)
Clarke, F.H., Ledyaev, Y.S., Stern, R.J., Wolenski, P.R.: Nonsmooth Analysis and Control Theory. Springer, New York (1998)
Daniilidis, A., Georgiev, P.: Approximate convexity and submonotonicity. J. Math. Anal. Appl. 291 (1), 292–301 (2004)
Djelassi, H., Mitsos, A.: A hybrid discretization algorithm with guaranteed feasibility for the global solution of semi-infinite programs. J Glob Optim. 68, 227–253 (2017)
Floudas, C.A., Stein, O.: The adaptive convexification algorithm: a feasible point method for semi-infinite programming. SIAM J. Optim. 18, 1187–1208 (2007)
Fuduli, A., Gaudioso, M., Giallombardo, G., Miglionico, G.: A partially inexact bundle method for convex semi-infinite minmax problems. Commun. Nonlinear Sci. Numer. Simul. 21, 172–180 (2015)
Gustafson, S.-Å., Kortanek, K.: Semi-infinite programming and applications. In: Mathematical Programming The State of the Art, pp 132–157. Springer, Berlin (1983)
Hare, W., Sagastizábal, C.: A redistributed proximal bundle method for nonconvex optimization. SIAM J. Optim. 20 (5), 2442–2473 (2010)
Hare, W., Sagastizábal, C., Solodov, M.: A proximal bundle method for nonsmooth nonconvex functions with inexact information. Comput. Optim. Appl. 63, 1–28 (2016)
Hettich, R.: An implementation of a discretization method for semi-infinite programming. Math. Program. 34 (3), 354–361 (1986)
Hettich, R., Kortanek, K.O.: Semi-infinite programming: theory, methods, and applications. Soc. Ind. Appl. Math. 35 (3), 380–429 (1993)
Hoseini Monjezi, N., Nobakhtian, S.: A filter proximal bundle method for nonsmooth nonconvex constrained optimization. J. Glob. Optim. (2020)
Hoseini Monjezi, N., Nobakhtian, S.: A new infeasible proximal bundle algorithm for nonsmooth nonconvex constrained optimization. Comput Optim Appl. 74, 443–480 (2019)
Karas, E., Ribeiro, A., Sagastizábal, C., Solodov, M.: A bundle-filter method for nonsmooth convex constrained optimization. Math. Program. 116, 297–320 (2009)
Karmitsa, N.: Test problems for large-scale nonsmooth minimization. Reports of the Department of Mathematical Information Technology, Series B, Scientific computing, No. B 4/2007 (2007)
Kiwiel, K.C.: Methods of Descent for Nondifferentiable Optimization, vol. 1133. Springer, Berlin, Heidelberg (2006)
Kortanek, K.O., No, H.: A central cutting plane algorithm for convex semi-infinite programming problems. SIAM J. Optim. 3 (4), 901–918 (1993)
López, M., Still, G.: Semi-infinite programming. Eur. J. Oper. Res. 180 (2), 491–518 (2007)
Lv, J., Pang, L.P., Meng, F.Y.: A proximal bundle method for constrained nonsmooth nonconvex optimization with inexact information. J. Global Optim. 70, 517–549 (2018)
Lv, J., Pang, L.P., Xu, N., Xiao, Z.H.: An infeasible bundle method for nonconvex constrained optimization with application to semi-infinite programming problems. Numer. Algorithms 80, 397–427 (2019)
Meng, F.Y., Pang, L.P., Lv, J., Wang, J.H.: An approximate bundle method for solving nonsmooth equilibrium problems. J. Global Optim. 68, 537–562 (2017)
Mifflin, R.: Semismooth and semiconvex functions in constrained optimization. SAIM J. Control Optim. 15 (6), 959–972 (1977)
Mitsos, A.: Global optimization of semi-infinite programs via restriction of the right-hand side. Optimization 60 (10-11), 1291–1308 (2011)
Mitsos, A., Djelassi, H.: A Test Set of Semi-Infinite Programs. Process Systems Engineering (AVT.SVT), RWTH Aachen University, Aachen, Germany. http://web.mit.edu/mitsos/www/pubs/siptestset.pdf (2016)
Mitsos, A., Lemonidis, P., Lee, C.K., Barton, P.I.: Relaxation-based bounds for semi-infinite programs. SIAM J. Optim. 19, 77–113 (2007)
Mitsos, A., Tsoukalas, A.: Global optimization of generalized semi-infinite programs via restriction of the right hand side. J. Glob. Optim. 61 (1), 1–17 (2015)
Pang, L.P., Lv, J., Wang, J.H.: Constrained incremental bundle method with partial inexact oracle for nonsmooth convex semi-infinite programming problems. Comput. Optim. Appl. 64 (2), 433–465 (2016)
Pang, L., Wu, Q., Wang, J., et al.: A discretization algorithm for nonsmooth convex semi-infinite programming problems based on bundle methods. Comput Optim Appl. 76, 125–153 (2020)
Polyak, B.T.: Introduction to Optimization. Optimization Software, Inc., Publications Division, New York (1987)
Rustem, B., Nguyen, Q.: An algorithm for the inequality-constrained discrete min-max problem. SIAM J. Optim. 8, 265–283 (1998)
Sagastizábal, C., Solodov, M.: An infeasible bundle method for nonsmooth convex constrained optimization without a penalty function or a filter. SIAM J. Optim. 16 (1), 146–169 (2005)
Shiu, T.J., Wu, S.Y.: Relaxed cutting plane method with convexification for solving nonlinear semi-infinite programming problems. Comput Optim Appl. 53, 91–113 (2012)
Spingarn, J.E.: Submonotone subdifferentials of Lipschitz functions. Trans. Am. Math. Soc. 264, 77–89 (1981)
Stein, O., Steuermann, P.: The adaptive convexification algorithm for semi-infinite programming with arbitrary index sets. Math. Program. 136, 183–207 (2012)
Still, G.: Discretization in semi-infinite programming: the rate of convergence. Math. Programm. 91 (1), 53–69 (2001)
Tang, C., Liu, S., Jian, J., et al.: A feasible SQP-GS algorithm for nonconvex, nonsmooth constrained optimization. Numer Algor. 65, 1–22 (2014)
Wang, S.X., Yuan, Y.X.: Feasible method for semi-infinite programs. SIAM J. Optim. 25 (4), 2537–2560 (2015)
Xu, M.W., Wu, S.Y., Ye, J.J.: Solving semi-infinite programs by smoothing projected gradient method. Comput. Optim. Appl. 59, 591–616 (2014)
Yang, Y., Pang, L., Ma, X., et al.: Constrained Nonconvex Nonsmooth Optimization via Proximal Bundle Method. J Optim Theory Appl. 163, 900–925 (2014)
Zhang, L.P., Wu, S.Y., López, M.A.: A new exchange method for convex semi-infinite programming. SIAM J. Optim. 20 (6), 2959–2977 (2010)
Funding
This work was supported by the Key Research and Development Projects of Shandong Province (NO. 2019GGX104089), and the Natural Science Foundation of Shandong Province (NO. ZR2019BA014).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix: 1: Test problems
Appendix: 1: Test problems
Problem 1
[23, 40] Dimension: n= 2, \(f(x) = 8|{x_{1}^{2}}-x_{2}|+(1-x_{1})^{2}\), \(g(x) = \max \limits \left \{\sqrt {2}x_{1},\ 2 x_{2}\right \}-1\), X = [− 2, 2]2, x0 = (1, 1)T.
Problem 2
[34] Dimension: n= 2, \(f(x) = \max \limits \left \{{{x_{1}^{2}}}+{{x_{2}^{4}}},\ (2-x_{1})^{2}+(2-x_{2})^{2},\ 2\exp (x_{2}-x_{1})\right \}\), \(g(x) = \max \limits \left \{-{{x_{1}^{4}}}-2{{x_{2}^{2}}}-1,\ 2 {{x_{1}^{3}}}-{{x_{2}^{2}}}-2.5\right \}\), X = [− 4, 4]2, x0 = (2, 2)T.
Problem 3
[34] Dimension: n= 2, \(f(x) = \max \limits \left \{{{x_{1}^{2}}}+{{x_{2}^{2}}}+x_{1} x_{2},\ -{{x_{1}^{2}}}-{{x_{2}^{2}}}-x_{1} x_{2},\ \sin \limits {x_{1}},\! -\sin \limits {x_{1}},\\ \cos \limits {x_{2}},\ -\cos \limits {x_{2}}\right \}\), \(g(x) = \max \limits \left \{-{{x_{1}^{4}}}-2{{x_{2}^{2}}}-1,\ 2 {{x_{1}^{3}}}-{{x_{2}^{2}}}-2.5\right \}\), X = [− 4, 4]2, x0 = (3, 1)T.
Problem 4
[34] Dimension: n= 2, \(f(x) = \max \limits \left \{{{x_{1}^{4}}}+{{x_{2}^{2}}},\ (2-x_{1})^{2}+(2-x_{2})^{2},\ 2\exp (x_{2}-x_{1})\right \}\), \(g(x) = \max \limits \left \{{{x_{1}^{2}}}-{{x_{2}^{2}}},\ -2{{x_{1}^{3}}}-{{x_{2}^{2}}}\right \}\), X = [− 4, 4]2, x0 = (0, 1)T.
Problem 5
[34] Dimension: n= 2, \(f(x) = \max \limits \left \{{{x_{1}^{2}}}+{{x_{2}^{2}}},\ (2-x_{1})^{2}+(2-x_{2})^{2},\ 2\exp (x_{2}-x_{1})\right \}\), \(g(x) = \max \limits \left \{{x_{1}}+{x_{2}}-2,\ -{{x_{1}^{2}}}-{{x_{2}^{2}}}+2.25\right \}\), X = [− 4, 4]2, x0 = (2.1, 1.9)T.
Problem 6
[34] Dimension: n= 2, \(f(x) = \max \limits \left \{10(x_{2}-{x_{1}^{2}}),\ 10({x_{1}^{2}}-x_{2}),\ 1-x_{1},\ x_{1}-1\right \}\), \(g(x) = \max \limits \left \{100{x_{1}^{2}}+{x_{2}^{2}}-101,\ 80{x_{1}^{2}}-{x_{2}^{2}}-79\right \}\), X = [− 4, 4]2, x0 = (− 1.2, 1)T.
Problem 7
[19] Dimension: n= 20, 50, 100, 200, \(f(x) =\max \limits \left \{{\sum }_{i=1}^{n-1}({x_{i}^{2}}+(x_{i+1}-1)^{2}+x_{i+1}-1)\right .\), \(~~~~~~~~~~~~~\left .{\sum }_{i=1}^{n-1}(-{x_{i}^{2}}-(x_{i+1}-1)^{2}+x_{i+1}+1)\right \}\), \(g(x) = {\sum }_{i=1}^{n-1}({x_{i}^{2}}+x_{i+1}^{2}+x_{i} x_{i+1}-2x_{i}-2x_{i+1}+1.0)\), X = [− 10, 10]n, x0 = ones(n, 1).
Problem 8
[19, 40] Dimension: n= 10, 50, 100, 200, 500, 1000, \(f(x) = {\sum }_{i=1}^{n-1}(-x_{i}+2({x_{i}^{2}}+x_{i+1}^{2}-1)+1.75|{x_{i}^{2}}+x_{i+1}^{2}-1|)\), \(g(x) = {\sum }_{i=1}^{n-2}((3-2 x_{i+1})x_{i+1}-x_{i}-2x_{i+2}+2.5)\), X = [− 10, 10]n, x0 = ones(n, 1).
Problem 9
[41] Dimension: n = 2, p = 1, \(f(x) = \frac {1}{3}{x_{1}^{2}}+{x_{2}^{2}}+\frac {1}{2}x_{1}\), \(g(x,t) = (1-{x_{1}^{2}} t^{2})^{2}-x_{1} t^{2}-{x_{2}^{2}}+x_{2}\), X = [− 2, 2]2, T = [0, 1], x0 = (− 1,− 1)T.
Problem 10
[28] Dimension: n = 3, p = 1, \(f(x) = \exp (x_{1})+\exp (x_{2})+\exp (x_{3})\), g(x,t) = 1/(1 + t2) − x1 − x2t − x3t2, X = [− 2, 2]3, T = [0, 1], x0 = (1, 1, 1)T.
Problem 11
[28] Dimension: n = 3, p = 1, f(x) = x1 + x2/2 + x3/3, \(g(x,t) = \exp (t-1)-x_{1} -x_{2} t-x_{3} t^{2}\), X = [− 2, 2]3, T = [0, 1], x0 = (1, 1, 1)T.
Problem 12
[41] Dimension: n = 3, p = 1, \(f(x) = {x_{1}^{2}}+{x_{2}^{2}}+{x_{3}^{2}}\), \(g(x,t) = x_{1} +x_{2}\exp (x_{3} t)+\exp (2t)-2\sin \limits (4t)\), X = [− 4, 2]3, T = [0, 1], x0 = (1, 1, 1)T.
Problem 13
[31, 32] Dimension: n = 3, p = 2, f(x) = |x1| + |x2| + |x3|, \(g(x,t) = x_{1}+x_{2}\exp (x_{3} t)-\exp (2x_{1} t)+\sin \limits (4 t)\), X = [− 1, 1]3, T = [0, 1], x0 = ones(3, 1).
Problem 14
[31, 32] Dimension: n = 3, p = 2, f(x) = |x1| + |x2| + |x3|, \(g(x,t) = x_{1}+x_{2}\exp (x_{3} t_{1})-\exp (2 t_{2})+\sin \limits (4 t_{1})\), X = [− 1, 1]3, T = [0, 1]2, x0 = ones(3, 1).
Problem 15
[31, 32] Dimension: n = 4, p = 2, f(x) = 1/2(|x1| + |x2| + |x3| + |x4|), \(g(x,t) = \sin \limits (t_{1} t_{2})-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} t_{1} t_{2}\), X = [− 4, 4]6, T = [0, 1]2, x0 = ones(4, 1).
Problem 16
Dimension: n = 6, p = 2, \(f(x) = {\sum }_{i=1}^{n-1}(-x_{i}+2({x_{i}^{2}}+x_{i+1}^{2}-1)+1.75|{x_{i}^{2}}+x_{i+1}^{2}-1|)\), \(g(x,t) = \sin \limits (t_{1} t_{2})-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\), X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 17
Dimension: n = 6, p = 2, \(f(x) = {\sum }_{i=1}^{n-1}(-x_{i}+2({x_{i}^{2}}+x_{i+1}^{2}-1)+1.75|{x_{i}^{2}}+x_{i+1}^{2}-1|)\), \(g(x,t) = (1+{t_{1}^{2}}+{t_{2}^{2}})^{2}-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\), X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 18
Dimension: n = 6, p = 2, \(f(x) = {\sum }_{i=1}^{n-1}\max \limits \left \{{x_{i}^{2}}+(x_{i+1}-1)^{2}+x_{i+1}-1,\ -{x_{i}^{2}}-(x_{i+1}-1)^{2}+x_{i+1}+1\right \}\), \(g(x,t) = \exp ({t_{1}^{2}}+{t_{2}^{2}})-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\), X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 19
Dimension: n = 6, p = 2, \(f(x) = {\sum }_{i=1}^{n-1}\max \limits \left \{{x_{i}^{2}}+(x_{i+1}-1)^{2}+x_{i+1}-1,\ -{x_{i}^{2}}-(x_{i+1}-1)^{2}+x_{i+1}+1\right \}\), \(g(x,t) = \sin \limits (t_{1} t_{2})-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\), X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 20
Dimension: n = 6, p = 2, \(f(x) = {\sum }_{i=1}^{n-1}\max \limits \left \{{x_{i}^{2}}+(x_{i+1}-1)^{2}+x_{i+1}-1,\ -{x_{i}^{2}}-(x_{i+1}-1)^{2}+x_{i+1}+1\right \}\), \(g(x,t) = (1+{t_{1}^{2}}+{t_{2}^{2}})^{2}-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\), X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 21
Dimension: n = 6, p = 2,
\(f(x)\! =\! \max \limits \left \{{\sum }_{i=1}^{n-1}({x_{i}^{2}}+(x_{i+1}-1)^{2}+x_{i+1}-1)\right .\), \(~~~~~~~~~~~~~~~~~\left .{\sum }_{i=1}^{n-1}(-{x_{i}^{2}}-(x_{i+1}-1)^{2}+x_{i+1}+1)\right \}\),
\(g(x,t) = \sin \limits (t_{1} t_{2})-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\),
X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 22
Dimension: n = 6, p = 2, \(f(x) =\max \limits \left \{\sum \limits _{i=1}^{n-1}({x_{i}^{2}}+(x_{i+1}-1)^{2}+x_{i+1}-1)\right .\), \(~~~~~~~~~~~~~\left .\sum \limits _{i=1}^{n-1}(-{x_{i}^{2}}-(x_{i+1}-1)^{2}+x_{i+1}+1)\right \}\),
\(g(x,t) = (1+{t_{1}^{2}}+{t_{2}^{2}})^{2}-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\),
X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Problem 23
Dimension: n = 6, p = 2,
\(f(x) = \max \limits _{1\leq i\leq n }\left \{h(-{\sum }_{i=1}^{n}x_{i}),\ h(x_{i})\right \}\), where \(h(y)=\ln (|y|+1),\ \forall y\in \mathbb {R}\),
\(g(x,t) = \sin \limits (t_{1} t_{2})-x_{1}-x_{2} t_{1}-x_{3} t_{2}-x_{4} {t_{1}^{2}}-x_{5} t_{1} t_{2}-x_{6} {t_{2}^{2}}\),
X = [− 4, 4]6, T = [0, 1]2, x0 = ones(6, 1).
Rights and permissions
About this article
Cite this article
Pang, LP., Wu, Q. A feasible proximal bundle algorithm with convexification for nonsmooth, nonconvex semi-infinite programming. Numer Algor 90, 387–422 (2022). https://doi.org/10.1007/s11075-021-01192-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11075-021-01192-9