Differential extraproximal method for finding an equilibrium in two-person saddle-point games (Q416804)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Differential extraproximal method for finding an equilibrium in two-person saddle-point games
scientific article

    Statements

    Differential extraproximal method for finding an equilibrium in two-person saddle-point games (English)
    0 references
    0 references
    0 references
    10 May 2012
    0 references
    The authors start with an economic game related to a project of investment in new technologies. The players are creditors and debtors. An equilibrium point defines the point of intersection of a demand curve and supply curve formed by debtors behavior and interest rate, respectively. If such an equilibrium does not exist the players can continue to bargain and the parties can agree after the process of communication and negotiation that the desired equilibrium has been found with some accuracy. The primary game is then transformed into the so-called two person saddle-point game by assuming that strategies of the players are saddle-points of Lagrange functions formulated for mathematical programming problems related to the optimization tasks of the players. Then the authors present relationships between the formulated saddle-point game and multiobjective optimization problem in which Pareto points are searched. The next step proposed in the paper is to transform the game into a form suitable for the use of the extraproximal method [the first author, Metody optimizatsii (Optimization methods), Moscow (2002)]. In this formulation the designed equilibrium becomes a fixed-point of the proximal operator. Finally to find the fixed-point the authors use the natural saddle-point method in the form of a Cauchy problem for the formulated system of ordinary differential equations. Since generally the system is unstable three versions of the differential extraproximal method for finding the solution are proposed and their features are discussed.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    saddle-point game
    0 references
    extraproximal method
    0 references
    Pareto optimality
    0 references