Abstract
In this paper, we study second-order differentiability properties of probability functions. We present conditions under which probability functions involving nonlinear systems and Gaussian (or Student) multi-variate random vectors are twice continuously differentiable. We provide an expression for their Hessian that can be useful in numerical methods for solving probabilistic constrained optimization problems.
Similar content being viewed by others
Notes
Since we do not assume complex structure on g, the general case reduces to the normal centered case, as follows. Let \(\tilde{\xi }\sim \mathcal {N}(\mu ,\Sigma )\) be a general multi-variate Gaussian random vector with mean \(\mu \) and covariance matrix \(\Sigma \). We define \(\tilde{g}(x,z) = g(x, Dz + \mu )\) and observe that \(\xi = D^{-1}(\tilde{\xi }- \mu ) \sim \mathcal {N}(0,R)\), where D is the diagonal matrix with \(D_{ii} = {\Sigma _{ii}}^{1/2}\). We see that \(\mathbb {P}[ g(x,\tilde{\xi }) \le 0] = \mathbb {P}[ \tilde{g}(x,\xi ) \le 0 ]\) and that neither convexity in the second argument of g, nor its continuous differentiability properties are perturbed by such a transformation.
References
Arnold, T., Henrion, R., Möller, A., Vigerske, S.: A mixed-integer stochastic nonlinear optimization problem with joint probabilistic constraints. Pacific J. Optim. 10, 5–20 (2014)
Bremer, I., Henrion, R., Möller, A.: Probabilistic constraints via SQP solver: application to a renewable energy management problem. Comput. Manag. Sci. 12, 435–459 (2015)
Deák, I.: Computing probabilities of rectangles in case of multinormal distribution. J. Stat. Comput. Simul. 26(1–2), 101–114 (1986)
Deák, I.: Subroutines for computing normal probabilities of sets—computer experiences. Ann. Oper. Res. 100, 103–122 (2000)
Genz, A., Bretz, F.: Computation of multivariate normal and t probabilities. Number 195 in Lecture Notes in Statistics. Springer, Dordrecht (2009)
Henrion, R., Möller, A.: Optimization of a continuous distillation process under random inflow rate. Computer Math. Appl. 45, 247–262 (2003)
Henrion, R., Möller, A.: A gradient formula for linear chance constraints under Gaussian distribution. Math. Oper. Res. 37, 475–488 (2012)
Henrion, R., Strugarek, C.: Convexity of chance constraints with independent random variables. Comput. Optim. Appl. 41, 263–276 (2008)
Henrion, R., Strugarek, C.: Convexity of chance constraints with dependent random variables: the use of copulae. In: Bertocchi, M., Consigli, G., Dempster, M.A.H. (eds.) Stochastic Optimization Methods in Finance and Energy: New Financial Products and Energy Market Strategies. International Series in Operations Research and Management Science, pp. 427–439. Springer, New York (2011)
Hunter, J.K., Nachtergaele, B.: Applied Analysis. World Scientific Publishing Company, Singapore (2001)
Izmailov, A.F., Solodov, M.V.: Newton-Type Methods for Optimization and Variational Problems. Springer Series in Operations Research and Financial Engineering. Springer, New York (2014)
Morgan, D.R., Eheart, J.W., Valocchi, A.J.: Aquifer remediation design under uncertainty using a new chance constraint programming technique. Water Resour. Res. 29, 551–561 (1993)
Naor, A., Romik, D.: Projecting the surface measure of the sphere of \(\ell _p^n\). Ann. I.H. Poincaré 39(2), 241–261 (2003)
Prékopa, A.: Stochastic Programming. Kluwer, Dordrecht (1995)
Prékopa, A.: Probabilistic programming. In: Ruszczyński, A., Shapiro, A. (eds.) Stochastic Programming. Handbooks in Operations Research and Management Science, vol. 10, pp. 267–351. Elsevier, Amsterdam (2003)
Shapiro, A., Dentcheva, D., Ruszczyński, A.: Lectures on Stochastic Programming. Modeling and Theory, MPS-SIAM series on optimization, vol. 9. SIAM and MPS, Philadelphia (2009)
Uryas’ev, S.: Derivatives of probability and integral functions: general theory and examples, 2nd edn. In: Floudas, C.A., Pardalos, P.M. (eds.) Encyclopedia of Optimization, pp. 658–663. Springer, New York (2009)
van Ackooij, W.: Decomposition approaches for block-structured chance-constrained programs with application to hydro-thermal unit commitment. Math. Methods Oper. Res. 80(3), 227–253 (2014)
van Ackooij, W.: Eventual convexity of chance constrained feasible sets. Optimization (J. Math. Program. Oper. Res) 64(5), 1263–1284 (2015)
van Ackooij, W., de Oliveira, W.: Level bundle methods for constrained convex optimization with various oracles. Comput. Optim. Appl. 57(3), 555–597 (2014)
van Ackooij, W., Henrion, R.: Gradient formulae for nonlinear probabilistic constraints with Gaussian and Gaussian-like distributions. SIAM J. Optim. 24(4), 1864–1889 (2014)
van Ackooij, W., Henrion, R., Möller, A., Zorgati, R.: On probabilistic constraints induced by rectangular sets and multivariate normal distributions. Math. Methods Oper. Res. 71(3), 535–549 (2010)
van Ackooij, W., Henrion, R., Möller, A., Zorgati, R.: Joint chance constrained programming for hydro reservoir management. Optim. Eng. 15, 509–531 (2014)
van Ackooij, W., Minoux, M.: A characterization of the subdifferential of singular Gaussian distribution functions. Set Valued Var. Anal. 23(3), 465–483 (2015)
van Ackooij, W., Sagastizábal, C.: Constrained bundle methods for upper inexact oracles with application to joint chance constrained energy problems. SIAM J. Optim. 24(2), 733–765 (2014)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
van Ackooij, W., Malick, J. Second-order differentiability of probability functions. Optim Lett 11, 179–194 (2017). https://doi.org/10.1007/s11590-016-1015-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11590-016-1015-7