Gradient-based optimisation of the conditional-value-at-risk using the multi-level Monte Carlo method
From MaRDI portal
Publication:6087955
Abstract: In this work, we tackle the problem of minimising the Conditional-Value-at-Risk (CVaR) of output quantities of complex differential models with random input data, using gradient-based approaches in combination with the Multi-Level Monte Carlo (MLMC) method. In particular, we consider the framework of multi-level Monte Carlo for parametric expectations and propose modifications of the MLMC estimator, error estimation procedure, and adaptive MLMC parameter selection to ensure the estimation of the CVaR and sensitivities for a given design with a prescribed accuracy. We then propose combining the MLMC framework with an alternating inexact minimisation-gradient descent algorithm, for which we prove exponential convergence in the optimisation iterations under the assumptions of strong convexity and Lipschitz continuity of the gradient of the objective function. We demonstrate the performance of our approach on two numerical examples of practical relevance, which evidence the same optimal asymptotic cost-tolerance behaviour as standard MLMC methods for fixed design computations of output expectations.
Recommendations
- Simulation optimization of risk measures with adaptive risk levels
- Robust Optimization of PDEs with Random Coefficients Using a Multilevel Monte Carlo Method
- Multilevel quasi-Monte Carlo for optimization under uncertainty
- MG/OPT and multilevel Monte Carlo for robust optimization of PDEs
- Stochastic quasi-gradient techniques in VaR-based ALM models
Cites work
- A fully asynchronous multifrontal solver using distributed dynamic scheduling
- A Quasi-Monte Carlo Method for Optimal Control Under Uncertainty
- Adaptive multilevel Monte Carlo simulation
- Adaptive sampling strategies for stochastic optimization
- Automated solution of differential equations by the finite element method. The FEniCS book
- Coherent measures of risk
- Lectures on Stochastic Programming
- Multifidelity preconditioning of the cross-entropy method for rare event simulation and failure probability estimation
- Multilevel Monte Carlo approximation of functions
- Multilevel Monte Carlo Path Simulation
- On optimal and data-based histograms
- Optimal control of convective FitzHugh-Nagumo equation
- Optimization of Convex Risk Functions
- Performance and scalability of the block low-rank multifrontal factorization on multicore architectures
- Portfolio optimization by minimizing conditional value-at-risk via nondifferentiable optimization
- Quantifying uncertain system outputs via the multilevel Monte Carlo method. I: Central moment estimation
- Riccati-based feedback control of the monodomain equations with the Fitzhugh-Nagumo model
- Risk-averse design of tall buildings for uncertain wind conditions
- Risk-averse PDE-constrained optimization using the conditional value-at-risk
- Robust Optimization of PDEs with Random Coefficients Using a Multilevel Monte Carlo Method
- Sample size selection in optimization methods for machine learning
- Simulating sensitivities of conditional value at risk
Cited in
(3)
This page was built for publication: Gradient-based optimisation of the conditional-value-at-risk using the multi-level Monte Carlo method
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6087955)
