Exploitation and Safety in General Sum Games

ثبت نشده
چکیده

We describe a method for an agent playing a generalsum normal form game to balance the rewards of exploiting a prediction of opponent behavior with the risks of being exploited by a self-interested opponent while guaranteeing a worst-case safety margin. Our algorithm, Restricted Stackelberg Response with Safety, calculates a probability distribution over the agent’s moves that balances those concerns. The probability distribution is generated by finding a modification of the Stackelberg response in a modified game which reflects the assumption that the opponent will behave according to the prediction with some probability, but may best-respond. We prove that the tradeoff provided by a Restricted Stackelberg Response between taking advantage of a prediction and avoiding exploitation can be computed. We show experimental results using Restricted Stackelberg Response with Safety in a generalsum game.

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

Identifying and Responding to Cooperative Actions in General-sum Normal Form Games

My primary research interest is social behavior of software agents acting in general-sum normal form games. In an environment with another agent, an agent needs to be able to identify if the opponent is hostile or cooperative by looking at the opponent’s actions, and respond appropriately. If the opponent is hostile, the agent should guard against it. Even for a self-interested agent, cooperati...

متن کامل

Safe Exploitation of Predictions of Opponent Behavior

Given a prediction of opponent behavior in a general-sum two-player normal form game, it is difficult to select a strategy that balances the opportunity to use the prediction to inform one’s action with the risk of becoming predictable. We propose Restricted Stackelberg Response with Safety (RSRS), a novel way of generating such a strategy. RSRS uses an r-safe Stackelberg equilibrium in a modif...

متن کامل

Safely Using Predictions in General-Sum Normal Form Games

It is often useful to predict opponent behavior when playing a generalsum two-player normal form game. However best-responding to an inaccurate prediction can lead to a strategy which is vulnerable to exploitation. This paper proposes a novel method, Restricted Stackelberg Response with Safety (RSRS), for an agent to select a strategy to respond to a prediction. The agent uses the confidence it...

متن کامل

A TRANSITION FROM TWO-PERSON ZERO-SUM GAMES TO COOPERATIVE GAMES WITH FUZZY PAYOFFS

In this paper, we deal with games with fuzzy payoffs. We proved that players who are playing a zero-sum game with fuzzy payoffs against Nature are able to increase their joint payoff, and hence their individual payoffs by cooperating. It is shown that, a cooperative game with the fuzzy characteristic function can be constructed via the optimal game values of the zero-sum games with fuzzy payoff...

متن کامل

Joint Production Games and Share Functions

Player i's payoff in a noncooperative game is generally expressed as a function of the vector of strategies of all players. However, in some games 'simply reducible games' the payoff of player i is a function of two arguments the strategy chosen by i, and the sum of the strategies of all players in the game. Cournot oligopoly, public good provision, costand surplus-sharing, and open access reso...

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 2014