Searcharxiv⌕ Search

arXiv subjects

Rainer Buckdahn

Publications and source records attributed to Rainer Buckdahn.

At least 19 recordsLinked to original sources
arXiv2024

Optimal control problems with generalized mean-field dynamics and viscosity solution to Master Bellman equation

We study an optimal control problem of generalized mean-field dynamics with open-loop controls, where the coefficients depend not only on the state processes and controls, but also on the joint law of them. The value function $V$ defined in a conventional way, but it does not satisfy the Dynamic Programming Principle (DPP for short). For this reason we introduce subtly a novel value function $\vartheta$, which is closely related to the original value function $V$, such that, a description of $\vartheta$, as a solution of a partial differential equation (PDE), also characterizes $V$. We establish the DPP for $\vartheta$. By using an intrinsic notion of viscosity solutions, initially introduced in Burzoni, Ignazio, Reppen and Soner [8] and specifically tailored to our framework, we show that the value function $\vartheta$ is a viscosity solution to a Master Bellman equation on a subset of Wasserstein space of probability measures. The uniqueness of viscosity solution is proved for coefficients which depend on the time and the joint law of the control process and the controlled process. Our approach is inspired by Buckdahn, Li, Peng and Rainer [7], and leads to a generalization of the mean-field PDE in [7] to a Master Bellman equation in the case of controls.

math.OC↗
arXiv2024

A Global Stochastic Maximum Principle for Mean-Field Forward-Backward Stochastic Control Systems with Quadratic Generators

Our paper is devoted to the study of Peng's stochastic maximum principle (SMP) for a stochastic control problem composed of a controlled forward stochastic differential equation (SDE) as dynamics and a controlled backward SDE which defines the cost functional. Our studies combine the difficulties which come, on one hand, from the fact that the coefficients of both the SDE and the backward SDE are of mean-field type (i.e., they do not only depend on the control process and the solution processes but also on their law), and on the other hand, from the fact that the coefficient of the BSDE is of quadratic growth in $Z$. Our SMP is novel, it extends in a by far non trivial way existing results on SMP.

math.OC↗
arXiv2023

Path-depending controlled mean-field coupled forward-backward SDEs. The associated stochastic maximum principle

In the present paper we discuss a new type of mean-field coupled forward-backward stochastic differential equations (MFFBSDEs). The novelty consists in the fact that the coefficients of both the forward as well as the backward SDEs depend not only on the controlled solution processes $(X_t,Y_t,Z_t)$ at the current time $t$, but also on the law of the paths of $(X,Y,u)$ of the solution process and the control process. The existence of the solution for such a MFFBSDE which is fully coupled through the law of the paths of $(X,Y)$ in the coefficients of both the forward and the backward equations is proved under rather general assumptions. Concerning the law, we just suppose the continuity under the 2-Wasserstein distance of the coefficients with respect to the law of $(X,Y)$. The uniqueness is shown under Lipschitz assumptions and the non anticipativity of the law of $X$ in the forward equation. The main part of the work is devoted to the study of Pontryagin's maximal principle for such a MFFBSDE. The dependence of the coefficients on the law of the paths of the solution processes and their control makes that a completely new and interesting criterion for the optimality of a stochastic control for the MFFBSDE is obtained. In particular, also the Hamiltonian is novel and quite different from that in the existing literature. Last but not least, under the assumption of convexity of the Hamiltonian we show that our optimality condition is not only necessary but also sufficient.

math.PR↗
arXiv2022

Mean field stochastic control under sublinear expectation

Our work is devoted to the study of Pontryagin's stochastic maximum principle for a mean-field optimal control problem under Peng's $G$-expectation. The dynamics of the controlled state process is given by a stochastic differential equation driven by a $G$-Brownian motion, whose coefficients depend not only on the control, the controlled state process but also on its law under the $G$-expectation. Also the associated cost functional is of mean-field type. Under the assumption of a convex control state space we study the stochastic maximum principle, which gives a necessary optimality condition for control processes. Under additional convexity assumptions on the Hamiltonian it is shown that this necessary condition is also a sufficient one. The main difficulty which we have to overcome in our work consists in the differentiation of the $G$-expectation of parameterized random variables. As particularly delicate it turns out to handle with the $G$-expectation of a function of the controlled state process inside the running cost of the cost function. For this we have to study a measurable selection theorem for set-valued functions whose values are subsets of the representing set of probability measures for the $G$-expectation.

math.OC↗
arXiv2021

Mean-field BDSDEs and associated nonlocal semi-linear backward stochastic partial differential equations

In this paper we investigate mean-field backward doubly stochastic differential equations (BDSDEs), i.e., BDSDEs whose driving coefficients also depend on the joint law of the solution process as well as the solution of an associated mean-field forward SDE. Unlike the pioneering paper on BDSDEs by Pardoux-Peng (1994), we handle a driving coefficient in the backward integral of the BDSDE for which the Lipschitz assumption w.r.t. the law of the solution is sufficient, without assuming that this Lipschitz constant is small enough. On the other hand, as the parameters $(x,P_\xi)$ and $(x,P_\xi,y)$ run an infinite-dimensional space, unlike Pardoux and Peng, we cannot apply Kolmogorov's continuity criterion to the value function $V(t,x,P_{\xi}):=Y_t^{t,x,P_{\xi}}$, while in the classical case studied in Pardoux-Peng the value function $V(t,x)=Y_t^{t,x}$ can be shown to be of class $C^{1,2}([0,T]\times\mathbb{R}^d)$, we have for our value function $V(t,x,P_{\xi})$ and its derivative $\partial_\mu V(t,x,P_{\xi},y)$ only the $L^2$-differentiability with respect to $x$ and $y$, respectively. Using a new method we prove the characterization of $V=(V(t,x,P_{\xi}))$ as the unique solution of the associated mean-field backward stochastic PDE.

math.PR↗
arXiv2021

A General Conditional McKean-Vlasov Stochastic Differential Equation

In this paper we consider a class of {\it conditional McKean-Vlasov SDEs} (CMVSDE for short). Such an SDE can be considered as an extended version of McKean-Vlasov SDEs with common noises, as well as the general version of the so-called {\it conditional mean-field SDEs} (CMFSDE) studied previously by the authors [1, 14], but with some fundamental differences. In particular, due to the lack of compactness of the iterated conditional laws, the existing arguments of Schauder's fixed point theorem do not seem to apply in this situation, and the heavy nonlinearity on the conditional laws caused by change of probability measure adds more technical subtleties. Under some structure assumptions on the coefficients of the observation equation, we prove the well-posedness of solution in the weak sense along a more direct approach. Our result is the first that deals with McKean-Vlasov type SDEs involving state-dependent conditional laws.

math.PR↗
arXiv2020

Derivative over Wasserstein spaces along curves of densities

In this paper, given any random variable $\xi$ defined over a probability space $(\Omega,\mathcal{F},Q)$, we focus on the study of the derivative of functions of the form $L\mapsto F_Q(L):=f\big((LQ)_{\xi}\big),$ defined over the convex cone of densities $L\in\mathcal{L}^Q:=\{ L\in L^1(\Omega,\mathcal{F},Q;\mathbb{R}_+):\ E^Q[L]=1\}$ in $L^1(\Omega,\mathcal{F},Q).$ Here $f$ is a function over the space $\mathcal{P}(\mathbb{R}^d)$ of probability laws over $\mathbb{R}^d$ endowed with its Borel $\sigma$-field $\mathcal{B}(\mathbb{R}^d)$. The problem of the differentiability of functions $F_Q$ of the above form has its origin in the study of mean-field control problems for which the controlled dynamics admit only weak solutions. Inspired by P.-L. Lions' results [18] we show that, if for given $L\in\mathcal{L}^Q$, $L'\mapsto F_{LQ}(L'):\mathcal{L}^{LQ}\rightarrow\mathbb{R}$ is differentiable at $L'=1$, the derivative is of the form $g(\xi)$, where $g:\mathbb{R}^d\rightarrow\mathbb{R}$ is a Borel function which depends on $(Q,L,\xi)$ only through the law $(LQ)_\xi$. Denoting this derivative by $\partial_1F((LQ)_\xi,x):=g(x),\, x\in\mathbb{R}^d$, we study its properties, and we relate it to partial derivatives, recently investigated in [6], and, moreover, in the case when $f$ restricted to the 2-Wasserstein space $\mathcal{P}_2(\mathbb{R}^d)$ is differentiable in P.-L. Lions' sense and $(LQ)_{\xi}\in\mathcal{P}_2(\mathbb{R}^d)$, we investigate the relation between the derivative with respect to the density of $F_Q(L)=f\big((LQ)_{\xi}\big)$ and the derivative of $f$ with respect to the probability measure. Our main result here shows that $\partial_x\partial_1F((LQ)_\xi,x)=\partial_\mu f((LQ)_\xi,x),\ x\in \mathbb{R}^d,$ where $\partial_\mu f((LQ)_\xi,x)$ denotes the derivative of $f:\mathcal{P}_2(\mathbb{R}^d)\rightarrow \mathbb{R}$ at $(LQ)_\xi$.

math.PR↗
arXiv2017

A Mean-field Stochastic Control Problem with Partial Observations

In this paper we are interested in a new type of {\it mean-field}, non-Markovian stochastic control problems with partial observations. More precisely, we assume that the coefficients of the controlled dynamics depend not only on the paths of the state, but also on the conditional law of the state, given the observation to date. Our problem is strongly motivated by the recent study of the mean field games and the related McKean-Vlasov stochastic control problem, but with added aspects of path-dependence and partial observation. We shall first investigate the well-posedness of the state-observation dynamics, with combined reference probability measure arguments in nonlinear filtering theory and the Schauder fixed point theorem. We then study the stochastic control problem with a partially observable system in which the conditional law appears nonlinearly in both the coefficients of the system and cost function. As a consequence the control problem is intrinsically "time-inconsistent", and we prove that the Pontryagin Stochastic Maximum Principle holds in this case and characterize the adjoint equations, which turn out to be a new form of mean-field type BSDEs.

math.PR↗
arXiv2016

Unexpected Default in an Information Based Model

This paper provides sufficient conditions for the time of bankruptcy (of a company or a state) for being a totally inaccessible stopping time and provides the explicit computation of its compensator in a framework where the flow of market information on the default is modelled explicitly with a Brownian bridge between 0 and 0 on a random time interval.

math.PR↗
arXiv2016

Mean-field SDE driven by a fractional Brownian motion and related stochastic control problem

We study a class of mean-field stochastic differential equations driven by a fractional Brownian motion with Hurst parameter $H\in(1/2,1)$ and a related stochastic control problem. We derive a Pontryagin type maximum principle and the associated adjoint mean-field backward stochastic differential equation driven by a classical Brownian motion, and we prove that under certain assumptions, which generalise the classical ones, the necessary condition for the optimality of an admissible control is also sufficient.

math.OC↗
arXiv2016

Peng's Maximum Principle for a Stochastic Control Problem Driven by a Fractional and a Standard Brownian Motion

We study a stochastic control system involving both a standard and a fractional Brownian motion with Hurst parameter less than 1/2. We apply an anticipative Girsanov transformation to transform the system into another one, driven only by the standard Brownian motion with coefficients depending on both the fractional Brownian motion and the standard Brownian motion. We derive a maximum principle and the associated stochastic variational inequality, which both are generalizations of the classical case.

math.OC↗
arXiv2016

Brownian Bridges on Random Intervals

The issue of giving an explicit description of the flow of information concerning the time of bankruptcy of a company (or a state) arriving on the market is tackled by defining a bridge process starting from zero and conditioned to be equal to zero when the default occurs. This enables to catch some empirical facts on the behavior of financial markets: when the bridge process is away from zero, investors can be relatively sure that the default will not happen immediately. However, when the information process is close to zero, market agents should be aware of the risk of an imminent default. In this sense the bridge process leaks information concerning the default before it occurs. The objective of this first paper on Brownian bridges on stochastic intervals is to provide the basic properties of these processes.

math.PR↗
arXiv2015

Fully nonlinear stochastic and rough PDEs: Classical and viscosity solutions

We study fully nonlinear second-order (forward) stochastic partial differential equations (SPDEs). They can also be viewed as forward path-dependent PDEs (PPDEs) and will be treated as rough PDEs (RPDEs) under a unified framework. We develop first a local theory of classical solutions and define then viscosity solutions through smooth test functions. Our notion of viscosity solutions is equivalent to the alternative one using semi-jets. Next, we prove basic properties such as consistency, stability, and a partial comparison principle in the general setting. When the diffusion coefficient is semi-linear (but the drift can be fully nonlinear), we establish a complete theory, including global existence and comparison principle. Our methodology relies heavily on the method of characteristics.

math.PR↗
arXiv2014

Generalized Hamilton-Jacobi-Bellman equations with Dirichlet boundary and stochastic exit time optimal control problem

We consider a kind of stochastic exit time optimal control problems, in which the cost function is defined through a nonlinear backward stochastic differential equation. We study the regularity of the value function for such a control problem. Then extending Peng's backward semigroup method, we show the dynamic programming principle. Moreover, we prove that the value function is a viscosity solution to the following generalized Hamilton-Jacobi-Bellman equation with Dirichlet boundary: \[ \left\{ \begin{array} [c]{l} \inf\limits_{v\in V}\left\{\mathcal{L}(x,v)u(x)+f(x,u(x),\nabla u(x) \sigma(x,v),v)\right\}=0, \quad x\in D,\medskip\\ u(x)=g(x),\quad x\in \partial D, \end{array} \right. \] where $D$ is a bounded set in $\mathbb{R}^{d}$, $V$ is a compact metric space in $\mathbb{R}^{k}$, and for $u\in C^{2}(D)$ and $(x,v)\in D\times V$, \[\mathcal{L}(x,v)u(x):=\frac{1}{2}\sum_{i,j=1}^{d}(\sigma\sigma^{\ast})_{i,j}(x,v)\frac{\partial^{2}u}{\partial x_{i}\partial x_{j}}(x) +\sum_{i=1}^{d}b_{i}(x,v)\frac{\partial u}{\partial x_{i}}(x). \]

math.PR↗
arXiv2014

Value in mixed strategies for zero-sum stochastic differential games without Isaacs condition

In the present work, we consider 2-person zero-sum stochastic differential games with a nonlinear pay-off functional which is defined through a backward stochastic differential equation. Our main objective is to study for such a game the problem of the existence of a value without Isaacs condition. Not surprising, this requires a suitable concept of mixed strategies which, to the authors' best knowledge, was not known in the context of stochastic differential games. For this, we consider nonanticipative strategies with a delay defined through a partition $π$ of the time interval $[0,T]$. The underlying stochastic controls for the both players are randomized along $π$ by a hazard which is independent of the governing Brownian motion, and knowing the information available at the left time point $t_{j-1}$ of the subintervals generated by $π$, the controls of Players 1 and 2 are conditionally independent over $[t_{j-1},t_j)$. It is shown that the associated lower and upper value functions $W^π$ and $U^π$ converge uniformly on compacts to a function $V$, the so-called value in mixed strategies, as the mesh of $π$ tends to zero. This function $V$ is characterized as the unique viscosity solution of the associated Hamilton-Jacobi-Bellman-Isaacs equation.

math.PR↗
arXiv2014

Differential games with asymmetric information and without Isaacs condition

We investigate a two-player zero-sum differential game with asymmetric information on the payoff and without Isaacs condition. The dynamics is an ordinary differential equation parametrised by two controls chosen by the players. Each player has a private information on the payoff of the game, while his opponent knows only the probability distribution on the information of the other player. We show that a suitable definition of random strategies allows to prove the existence of a value in mixed strategies. Moreover, the value function can be characterised in term of the unique viscosity solution in some dual sense of a Hamilton-Jacobi-Isaacs equation. Here we do not suppose the Isaacs condition which is usually assumed in differential games.

math.OC↗
arXiv2014

Stochastic Variational Inequalities on Non-Convex Domains

The objective of this work is to prove, in a first step, the existence and the uniqueness of a solution of the following multivalued deterministic differential equation: $dx(t)+\partial ^-\varphi (x(t))(dt)\ni dm(t),\ t>0$, $x(0)=x_0$, where $m:\mathbb{R}_+\rightarrow\mathbb{R}^d$ is a continuous function and $\partial^-\varphi$ is the Fr\'{e}chet subdifferential of a semiconvex function $\varphi$; the domain of $\varphi$ can be non-convex, but some regularities of the boundary are required. The continuity of the map $m\mapsto x:C([0,T];\mathbb{R}^{d})\rightarrow C([0,T] ;\mathbb{R}^{d})$, which associate the input function $m$ with the solution $x$ of the above equation, as well as tightness criteria allow to pass from the above deterministic case to the following stochastic variational inequality driven by a multi-dimensional Brownian motion: $X_t+K_t = \xi+\int_0^t F(s,X_{s})ds + \int_0^t G(s,X_s) dB_s,\; t\geq0$, $\;$ with $dK_{t}(\omega)\in\partial^-\varphi( X_t (\omega))(dt)$.

math.DS↗