Odile POURTALLIER and Mabel TIDBALL N ̊ 2942 July 1996

نویسندگان

  • Odile POURTALLIER
  • Mabel TIDBALL
چکیده

We consider the approximation of a class of diierential games with target by stochastic games. We use Kruzkov transformation to obtain discounted costs. The approximation is based on a space discretization of the state space and leads to consider the value function of the diierential game as the limit of the value function of a sequence of stochastic games. To prove the convergence, we use the notion of viscosity solution for partial diierential equations. This allows us to make assumptions only on the continuity of the value function and not on its diierentiability. This technique of proof has been used before by M. Bardi, M. Falcone and P. Soravia for another kind of discretization. Under the additional hypothesis that the value function is Lipschitz continuous, we prove that the rate of convergence of this scheme is of order p h where h is the space parameter of discretization. Some numerical experiments are presented in order to test the algorithm for a problem with discontinuous solution. Approximation de la fonction valeur pour une classe de jeux diiirentiels avec cible RRsumm : On considdre l'approximation d'une classe de jeux diiirentiels avec cible par des jeux stochastiques. On utilise la transformation de Kruzkov pour obtenir des coots avec taux d'actualisation. Le schhma d'approximation utiliss est bass sur la discrrtisation de l'espace d''tat et revient considdrer la fonction Valeur du jeu diiirentiel comme la limite des fonctions Valeur d'une suite de jeux stochastiques. Pour prouver la convergence du schhma, on utilise la notion de solution de viscositt pour les quations aux derives partielles du premier ordre. Cela nous permet de restreindre les hypothhses sur la fonction Valeur la continuitt, sans se prroccuper de sa diiirentialitt. Les techniques de preuve ont tt ddjj utilisses par M. Bardi, M. Falcone et P. Soravia pour un autre schhma de discrrtisation. Sous l'hypothhse additionnelle de Lipschitzianitt de la fonction Valeur, on prouve que la vitesse de convergence du shhma tudii est de l'ordre de p h oo h est le parammtre de discrrtisation de l'espace. On prrsente quelques expriences nummriques pour tester l'algorithme dans un cas oo la fonction valeur est non continue.

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Continuity of Optimal Values and Solutions for Control of Markov Chains with Constraints

We consider in this paper constrained Markov decision processes. This type of control model has many applications in telecommunications and other fields [E. Altman and A. Shwartz, IEEE Trans. Automat. Control, 34 (1989), pp. 1089–1102, E. A. Feinberg and M. I. Reiman, Probab. Engrg. Inform. Sci., 8 (1994), pp. 463–489, A. Hordijk and F. Spieksma, Adv. in Appl. Probab., 21 (1989), pp. 409–431, A...

متن کامل

Approximations in Dynamic Zero-sum Games, Ii Approximations in Dynamic Zero-sum Games, Ii

We pursue in this paper our study of approximations of values and-saddle-point policies in dynamic zero-sum games. After extending the general theorem for approximation, we study zero-sum stochastic games with countable state space, and non-bounded immediate reward. We focus on the expected average payoo criterion. We use some tools developed in the rst paper, to obtain the convergence of the v...

متن کامل

Efficiency inducing taxation for polluting oligopolists: the irrelevance of privatization

This paper studies the optimal environmental policy in a mixed market when pollution accumulates over time. Specifically, we assume quantity competition between several private firms and one partially privatized firm. The optimal emission tax is shown to be independent of the weight the privatized firm puts on social welfare. The optimal tax rule, the accumulated stock of pollution, firms' prod...

متن کامل

Avoiding Paradoxes in Routing Games

Strange behavior may occur in networks due to the non-cooperative nature of decision making, when the latter are taken by individual agents. In particular, the well known Braess paradox illustrates that when upgrading a network by adding a link, the resulting equilibrium may exhibit larger delays for all users. We present here some guidelines to avoid the Braess paradox when upgrading a network...

متن کامل

Rate of Convergence of a Numerical Procedure for Impulsive Control Problems

In this paper we consider a deterministic impulsive control problem. We discretize the Hamilton-Jacobi-Bellman equation satissed by the optimal cost function and we obtain discrete solutions of the problem. We give an explicit rate of convergence of the approximate solutions to the solution of the original problem. We consider the optimal switching problem as a special case of impulsive control...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1996