Category Rapports

Avr
2020

A Principal-Agent approach to study Capacity Remuneration Mechanisms - Clémence Alasseur, Heythem Farhat and Marcelo Saguan

We propose to study electricity capacity remuneration mechanism design through a Principal-Agent approach. The Principal represents the aggregation of electricity consumers (or a representative entity), subject to the physical risk of shortage, and the Agent represents the electricity capacity owners, who invest in capacity and produce electricity to satisfy consumers’ demand, and are subject to financial risks. Following the methodology of Cvitanic et al. (2017), we propose an optimal contract, Read more [...]

Déc
2019

Numerical resolution of McKean-Vlasov FBSDEs using neural networks - Maximilien GERMAIN, Joseph MIKAEL, and Xavier WARIN

We propose several algorithms to solve McKean-Vlasov Forward Backward Stochastic Differential Equations. Our schemes rely on the approximating power of neural networks to estimate the solution or its gradient through minimization problems. As a consequence, we obtain methods able to tackle both mean field games and mean field control problems in high dimension. We analyze the numerical behavior of our algorithms on several examples including non linear quadratic models.

Juil
2019

Neural networks-based backward scheme for fully nonlinear PDEs - H. Pham, X. Warin

We propose a numerical method for solving high dimensional fully nonlinear partial differential equations (PDEs). Our algorithm estimates simultaneously by backward time induction the solution and its gradient by multi-layer neural networks, through a sequence of learning problems obtained from the minimization of suitable quadratic loss functions and training simulations. This methodology extends to the fully non- linear case the approach recently proposed in (Huré, Pham, Warin, 2019) for semi-linear Read more [...]

Juin
2019

Efficient Volatility Estimation in a Two-factor Model - O. Féron, P. Gruet, and M. Hoffmann

We statistically analyse a multivariate HJM diffusion model with stochastic volatility. The volatility process of the first factor is left totally unspecified while the volatility of the second factor is the product of an unknown process and an exponential function of time to maturity. This exponential term includes some real parameter measuring the rate of increase of the second factor as time goes to maturity. From historical data, we efficiently estimate the time to maturity parameter in the sense Read more [...]

Avr
2019

Simulation of fuel poverty in France - Corinne Chaton, Alexandre Gouraud.

The assessment of fuel poverty in mainland France is based mainly on data provided by the French national housing survey (ENL). However, the last two surveys date from 2006 and 2014. To understand the change in the number of fuel poverty households, we have developed a micro simulation tool that takes into account the three predominant factors in the notion of fuel poverty, that is, household resources, energy prices and dwelling quality. Our tool includes three multiple linear models for estimating Read more [...]

Avr
2019

Avoiding Fuel Poverty through Insurance -  Corinne Chaton

Twenty percent of French non-fuel poor households will fall into fuel poverty. The existence of energy insurance can reduce this percentage. This article focuses on non-fuel poor households that can buy insurance that provides a basic level of energy for one year after a significant loss of income. A model of household willingness to pay for energy insurance is proposed. Several simulations are performed with French data. Given the values of the utility function parameters and the energy prices, Read more [...]

Fév
2019

Some machine learning schemes for high-dimensional nonlinear PDEs - C. HURE, H. PHAM, X. WARIN

We propose new machine learning schemes for solving high dimensional nonlinear partial differential equations (PDEs). Relying on the classical backward stochastic differential equation (BSDE) representation of PDEs, our algorithms estimate simultaneously the solution and its gradient by deep neural networks. These approximations are performed at each time step from the minimization of loss functions de ned recursively by backward induction. The methodology is extended to variational inequalities Read more [...]

Fév
2019

Untangling systemic risk in financialized commodity markets - Julien Ling.

Systemic risk is a multifaceted concept that is of crucial importance for regulators. In order to ensure financial stability, they need to properly assess this risk, preventing financial shocks from affecting the real economy. In this study, we evaluate the extent to which the financialization of commodity markets contributes to systemic risk. We consider a system consisting of both commodity futures and financial markets in a sparse Vector AutoRegression (VAR) framework. It allows to distinguish Read more [...]

Jan
2019

Deep neural networks algorithms for stochastic control problems on finite horizon, Part 2: numerical applications - A. Bachouch, C. Huré, N. Langrené, H. Pham

This paper presents several numerical applications of deep learning-based algorithms that have been analyzed in [11]. Numerical and comparative tests using TensorFlow illustrate the performance of our different algorithms, namely control learning by performance iteration (algorithms NNcontPI and ClassifPI), control learning by hybrid iteration (algorithms Hybrid-Now and Hybrid-LaterQ), on the 100-dimensional nonlinear PDEs examples from [6] and on quadratic Backward Stochastic Differential equations Read more [...]

Jan
2019

Deep neural networks algorithms for stochastic control problems on finite horizon, part I : convergence analysis - C. Hure, H. Pham, A. Bachouch and N. Langrené

This paper develops algorithms for high-dimensional stochastic control problems based on deep learning and dynamic programming (DP). Diffrently from the classical approximate DP approach, we rst approximate the optimal policy by means of neural networks in the spirit of deep reinforcement learning, and then the value function by Monte Carlo regression. This is achieved in the DP recursion by performance or hybrid iteration, and regress now or later/quantization methods from numerical probabilities. Read more [...]

Page 1 of 10