A comparison of sample-based Stochastic Optimal Control methods

TL;DR

This paper compares scenario tree and particle-based methods for stochastic optimal control, highlighting the exponential growth in scenarios needed for scenario trees versus the particle method's scalability with time horizon, but noting challenges with high-dimensional states.

Contribution

It introduces a performance metric (MSE) for evaluating scenario-based methods and demonstrates how particle methods can overcome horizon-related complexity in stochastic control problems.

Findings

Scenario trees require exponentially more scenarios as the time horizon increases.

Particle methods maintain a constant number of scenarios regardless of the time horizon.

High-dimensional state spaces still pose significant challenges for particle methods.

Abstract

In this paper, we compare the performance of two scenario-based numerical methods to solve stochastic optimal control problems: scenario trees and particles. The problem consists in finding strategies to control a dynamical system perturbed by exogenous noises so as to minimize some expected cost along a discrete and finite time horizon. We introduce the Mean Squared Error (MSE) which is the expected $L^2$-distance between the strategy given by the algorithm and the optimal strategy, as a performance indicator for the two models. We study the behaviour of the MSE with respect to the number of scenarios used for discretization. The first model, widely studied in the Stochastic Programming community, consists in approximating the noise diffusion using a scenario tree representation. On a numerical example, we observe that the number of scenarios needed to obtain a given precision grows exponentially with the time horizon. In that sense, our conclusion on scenario trees is equivalent to the one in the work by Shapiro (2006) and has been widely noticed by practitioners. However, in the second part, we show using the same example that, by mixing Stochastic Programming and Dynamic Programming ideas, the particle method described by Carpentier et al (2009) copes with this numerical difficulty: the number of scenarios needed to obtain a given precision now does not depend on the time horizon. Unfortunately, we also observe that serious obstacles still arise from the system state space dimension.