Title:Social Optimization in Noncooperative Games under Central Regulation

Abstract:This paper proposes a novel optimization problem building on noncooperative games under central regulation, which can be formulated as a bilevel structure. In the low-level, each player competes to minimize its own cost function that depends not only on the strategies of all players, but also on an intervention decision of the central regulator, while the central regulator located at the high-level attempts to achieve the social optimum, that is, to minimize the sum of cost functions of all players through an adjustable intervention decision. In this setting, under the intervention of the central regulator, the low-level players perform in a noncooperative game and aim to seek the Nash equilibrium, which indeed is related with the regulator's decision. Meanwhile, the objective of the regulator is to choose a decision such that the social cost, i.e., the sum of cost functions of all players is minimum. This formulated bilevel social optimization problem is proven to be constrained, nonconvex and nonsmooth. To address this intricate problem, an inexact zeroth-order algorithm is developed by virtue of the smoothing techniques, allowing for the Nash equilibrium of the low-level game to be computed in an inexact manner. Levering the properties of smoothing techniques, it is rigorously shown that the devised algorithm achieves a sublinear convergence rate for computing a stationary point of a related optimization problem with a smoothed objective. Moreover, the sublinear convergence rate in the scenario where the exact equilibrium of the low-level game is available is also discussed. Finally, numerical simulations are conducted to demonstrate the efficiency of theoretical findings.