Finite-Agent Stochastic Differential Games on Large Graphs: I. The Linear-Quadratic Case

The subject of this paper is a non-autonomous linear quadratic case of a differential game model with continuous updating. This class of differential games is essentially new where it is assumed that, at each time instant, players have or use information about the game structure defined on a closed time interval with a fixed duration. During the interval information about motion equations and payoff functions of players updates. It is non-autonomy that simulates this effect of updating information. A linear quadratic case for this class of games is particularly important for practical problems arising in the engineering of human-machine interaction. Here we define the Nash equilibrium as an optimality principle and present an explicit form of Nash equilibrium for the linear quadratic case. Also, the case of dynamic updating for the linear quadratic differential game is studied and uniform convergence of Nash equilibrium strategies and corresponding trajectory for a case of continuous updating and dynamic updating is demonstrated.

The subject of this paper is a linear quadratic case of a differential game model with continuous updating. This class of differential games is essentially new, there it is assumed that at each time instant, players have or use information about the game structure defined on a closed time interval with a fixed duration. As time goes on, information about the game structure updates. Under the information about the game structure we understand information about motion equations and payoff functions of players. A linear quadratic case for this class of games is particularly important for practical problems arising in the engineering of human-machine interaction. The notion of Nash equilibrium as an optimality principle is defined and the explicit form of Nash equilibrium for the linear quadratic case is presented. Also, the case of dynamic updating for the linear quadratic differential game is studied and uniform convergence of Nash equilibrium strategies and corresponding trajectory for a case of continuous updating and dynamic updating is demonstrated.

Necessary/sufficient conditions for Pareto optimality in finite horizon mean-field type stochastic differential game

The main aim of this work is to develop a systematic approach for dealing with differential games with limited communication. To this end a differential game with limited communication is considered. The communication topology is described by a directed graph. The main components characterising the differential game with limited communication are identified before the resulting game is formally defined. Sufficient conditions to solve the problem are identified both in the general nonlinear case and in the linear-quadratic case. A numerical example illustrating the theoretical approach and results is presented. Finally, several directions for further developments are identified.

This paper extends the differential game exhaustible resource model of Jo;rgensen (1985) (Steffen Jorgensen, An Exponential Differential game Which Admits a Simple Nash Solution, J. Optim. Theory Appl., Vol. 45, No.3, March 1985, 383–396) by including a growth function for the stock of resource. Then the problem becomes one of managing a renewable resource, for example, a fishery. The model employs nonclassical assumptions on harvest rates and utility functions; these are supposed to be convex in effort. Hence, increasing returns to scale are assumed. From a theoretical point of view, the dynamic game model has an interesting feature. We show that it is possible to determine explicitly a feedback Nash equilibrium in effort rates by solving the Hamilton-Jacobi-Bellman equations for the value functions. This has only been possible in rather few cases: the linear-quadratic case and in some games of a simple (or special) structure. In other cases the dynamic programming approach has been used to find degenerate closed-loop solutions, i.e. closed-loop strategies that are independent of the state variable (see e.g. Jorgensen (1985)). In the present article, however, the Nash solutions depend explicitely on the state. When applied to the problem of optimal fishery management the closed-loop Nash strategies become, under certain circumstances, cyclic (periodic, pulsation).

This paper deals with the min-max and min-min Stackelberg strategies in the case of a closed-loop information structure. Two-player differential one-single stage games are considered with one leader and one follower. We first derive necessary conditions for the existence of the follower to characterize the best response set of the follower and to recast it, under weak assumptions, to an equivalent and more convenient form for expressing the constraints of the leader’s optimization problem. Under a standard strict Legendre condition, we then derive optimality necessary conditions for the leader of both min-max and min-min Stackelberg strategies in the general case of nonlinear criteria for finite time horizon games. This leads to an expression of the optimal controls along the associated trajectory. Then, using focal point theory, the necessary conditions are also shown to be sufficient and lead to cheap control. The set of initial states allowing the existence of an optimal trajectory is emphasized. The linear-quadratic case is detailed to illustrate these results.

Necessary and sufficient conditions for Pareto optimality of the stochastic systems in finite horizon

Strategic incentives in dynamic duopoly

The class of differential games with continuous updating is quite new, there it is assumed that at each time instant, players use information about the game structure (motion equations and payoff functions of players) defined on a closed time interval with a fixed duration. As time goes on, information about the game structure updates. A linear-quadratic case for this class of games is particularly important for practical problems arising in the engineering of human-machine interaction. In this paper, it is particularly interesting that the open-loop strategies are used to construct the optimal ones, but subsequently, we obtain strategies in the feedback form. Using these strategies the notions of Shapley value and Nash equilibrium as optimality principles for cooperative and non-cooperative cases respectively are defined and the optimal strategies for the linear-quadratic case are presented.

Strategic incentives in dynamic duopoly*1

Pareto-based Stackelberg differential game for stochastic systems with multi-followers

This paper constructs a non-cooperative/cooperative stochastic differential game model to prove that the optimal strategies trajectory of agents in a system with a topological configuration of a Multi-Local-World graph would converge into a certain attractor if the system’s configuration is fixed. Due to the economics and management property, almost all systems are divided into several independent Local-Worlds, and the interaction between agents in the system is more complex. The interaction between agents in the same Local-World is defined as a stochastic differential cooperative game; conversely, the interaction between agents in different Local-Worlds is defined as a stochastic differential non-cooperative game. We construct a non-cooperative/cooperative stochastic differential game model to describe the interaction between agents. The solutions of the cooperative and non-cooperative games are obtained by invoking corresponding theories, and then a nonlinear operator is constructed to couple these two solutions together. At last, the optimal strategies trajectory of agents in the system is proven to converge into a certain attractor, which means that strategies trajectory are certainty as time tends to infinity or a large positive integer. It is concluded that the optimal strategy trajectory with a nonlinear operator of cooperative/non-cooperative stochastic differential game between agents can make agents in a certain Local-World coordinate and make the Local-World payment maximize, and can make the all Local-Worlds equilibrated; furthermore, the optimal strategy of the coupled game can converge into a particular attractor that decides the optimal property.

Stochastic differential games are considered in a non-Markovian setting. Typically, in stochastic differential games the modulating process of the diffusion equation describing the state flow is taken to be Markovian. Then Nash equilibria or other types of solutions such as Pareto equilibria are constructed using Hamilton--Jacobi--Bellman (HJB) equations. But in a non-Markovian setting the HJB method is not applicable. To examine the non-Markovian case, this paper considers the situation in which the modulating process is a fractional Brownian motion. Fractional noise calculus is used for such models to find the Nash equilibria explicitly. Although fractional Brownian motion is taken as the modulating process because of its versatility in modeling in the fields of finance and networks, the approach in this paper has the merit of being applicable to more general Gaussian stochastic differential games with only slight conceptual modifications. This work has applications in finance to stock price modeling which incorporates the effect of institutional investors, and to stochastic differential portfolio games in markets in which the stock prices follow diffusions modulated with fractional Brownian motion.

We study a class of linear-quadratic stochastic differential games in which each player interacts directly only with its nearest neighbors in a given graph. We find a semiexplicit Markovian equilibrium for any transitive graph, in terms of the empirical eigenvalue distribution of the graph’s normalized Laplacian matrix. This facilitates large-population asymptotics for various graph sequences, with several sparse and dense examples discussed in detail. In particular, the mean field game is the correct limit only in the dense graph case, that is, when the degrees diverge in a suitable sense. Although equilibrium strategies are nonlocal, depending on the behavior of all players, we use a correlation decay estimate to prove a propagation of chaos result in both the dense and sparse regimes, with the sparse case owing to the large distances between typical vertices. Without assuming the graphs are transitive, we show also that the mean field game solution can be used to construct decentralized approximate equilibria on any sufficiently dense graph sequence.

In this chapter, we will deal with zero-sum two-player time-homogeneous stochastic differential games and viscosity solutions of the Isaacs equations arising from such games, via the dynamic programming principle.In Sect. 4.1, we are concerned with basic concepts and definitions and we introduce stochastic differential games, referring to (Controlled MarkovProcesses and viscosity solutions, 2nd edn. Springer, New York 2006), XI. Then, using a semi-discretization argument, we study the DPP for lower- and upper-value functions in Sect. 4.2. In Sect. 4.3, we will consider the Isaacs equations, via semigroups related to DPP. In Sect. 4.4, we consider a link between stochastic controls and differential games via risk sensitive controls.