Game theory applications computer science

It is the usual case that the central controlling for the wireless networks requires much information exchanges and coordination, and sometimes, it causes a huge overhead. In order to make the network more extensible, the distributed algorithm is required.

However, the rational or selfish nodes tend to optimize their own payoff without considering the social performance in the absence of the central controller. Therefore, the existing centralized schemes are no longer suitable for such case, whereas the game theory has the superiority in controlling the nodes in a distributed way.

Besides, sometimes, it is expected that the future wireless networks can support a variety of services with diverse quality of service QoS requirements. For instance, an application mixing of delay-sensitive and delay-tolerant requirements would coexist in the same network.

For instance, the voice applications pay more attention to the network delay, whereas the data transmission requires more bandwidth. Therefore, it is a main challenge to exploit the optimal overall performance for the network. The main contribution of the chapter lies as follows: the basic concept of the game theory and the three main components were introduced first. Then, the concept of wireless network was presented, and the main problems existing in the media access control level or the network level were also pointed out.

The superiority of the game theory in dealing with the conflict and cooperation was stated. Finally, the applications of the game theory in the wireless network were elaborated in detail. Game theory, which was presented in , is a theory concerning the decision-making. It gives some guides to the participants who face a dilemma whether to cooperate or conflict so as to obtain the maximum returns. In the game theory, an important concept Nash Equilibrium was proposed in which has promoted the research of the noncooperative game.

When a game model reaches the Nash Equilibrium, it means that any players can impossibly obtain more favorable utility via other actions. The game theory is a mathematical tool for analyzing the interactions of two or more decision-makers. The game theory is capable of stimulating the players to cooperate with each other to achieve a desirable goal.

Usually, a game model consists of three main components: the player set, the strategy set, and the utility function of each player. As for the wireless networks, the nodes lie in the same network segment constitute the set of the players. The strategy set consists of the choice which is made by the nodes when deciding whether to relay messages for others or not. The utility function should be designed carefully to stimulate the players to cooperate with each other to achieve a considerable overall goal.

It is worth noting that in the game theory, any actions taken by the user may affect the performance of others in the same network segment. The classical game theory bases on the assumption that all the players are perfectly rational. The prediction about the game is only agreement with the actual results when each player has perfect rationality. Nevertheless, this demand cannot always be met owing to some practical reasons. For instance, on account of the energy constraint, not every player is acquainted with the information of others.

Besides, the individual differences in intelligence and learning capacity also lead to the differences in the rational level. The game theoretic can be also divided into two main categories: the complete information game and the incomplete information game. For example, the incomplete game model can be applied to the problem of jamming for the wireless networks. There is also a kind of game model named the evolutionary game theory. The evolutionary game theory can be applied to the situation where each player is of limited rationality.

It was firstly introduced by Maynard Smith in [ 7 ]. Its development dues to the efforts which aim at explaining the evolution of genetically determined social behavior in the biological science. As we all know, in the real network environment, the assumption that all the players should be rational enough to determine their decisions is obviously not always satisfied.

In the wireless sensor networks, the nodes usually have limited rationality. So, the evolutionary game theory can be utilized to solve some issues in the wireless sensor networks. A wireless network is a kind of system which consists of a number of nodes communicating with each other via wireless data connection.

It is usually implemented via radio communications. The communication without cable can reduce the cost of deployment and maintenance, therefore the wireless network has gained a lot of applications. Examples of the wireless networks include the wireless local area networks WLAN , the wireless sensor networks WSNs , the ad hoc network, and the satellite communication networks. In the wireless communication scenario, a large number of nodes compete with each other for the common resource, such as the wireless channel, the bandwidth, etc.

Therefore, the data transmission follows the hop-by-hop transmission pattern. However, not all the nodes are willing to relay the data for others owing to the energy or bandwidth consumption for relaying data. Sometimes, the nodes tend to struggle with each other for the limited resource; the network capacity is reduced when it happens.

In the worst case, the data collision happens and it leads to the packet loss. The packet loss results in the decline of the network performance, such as the extension of the network delay.

So, how to stimulate the nodes to cooperative with each other so as to improve the network performance is a problem facing the wireless network. The wireless sensor networks WSNs is a kind of wireless network which consists of a huge number of tiny sensor nodes.

Therefore, it is impossible or unpractical to recharge the sensor nodes. When the portion of the energy-exhausting nodes reaches a certain threshold, the network partition generates.

For some applications, the network partition means the termination of the network life span. In order to extend the network lifetime, the energy efficiency should be improved. In general, the energy efficiency includes twofold, namely the minimization of energy consumption and the energy consumption equilibrium. The cooperation among the sensor nodes can improve the energy efficiency. Therefore, some incentive strategies should be designed to promote the cooperation among the nodes.

In conclusion, there are two main problems existing in the wireless networks, namely the network performance and the energy efficiency for the WSNs. The game theory has gained wide applications in improving the network performance and the energy efficiency. The state of the art in the applications of the game theory in the wireless networks was detailed in the chapter.

Different from the traditional local area networks LAN , the media access control for the wireless networks is more complicated owing to the openness of the media. Any node can get access to the media as long as it lies in the transmission range of another node. If two nodes which lie in their transmission range send data at the same time, the data collision happens.

The data collision has a bad influence on the network performance. Usually, the network performance is evaluated by the throughput, the packet loss ratio, the network delay, and the network delay jitters.

The data collision results in the packet loss and the decline of the network capacity, sometimes even the termination of the network life span. So, it is crucial to avoid the data collision in order to improve the network performance. In a distributed wireless system, a huge number of network nodes behave cooperative toward a common goal, such as environmental monitoring, emergency rescue, enemy tracking, and so on.

In such a scenario, how to attain mutual cooperation is an important scheme. Sometimes, not all of the nodes are willing to cooperate because it consumes much resource to relay messages for others.

For some extreme case, the task may be hardly to be completed. Recently, a lot of works have emerged concerning the network performance and they are introduced in detail in this section. It has been proven in the recent literature that the proper pricing techniques can be deployed among a number of users to achieve various resource allocation policies.

In the wireless relay networks, the relay nodes have no incentives to relay messages for the other users without an appropriate compensation mechanism, since it leads to the energy exhaust or the decline of the network capacity. So, the pricing mechanism provides a useful scheme that reimburses the relay nodes for using their resources by making some payoff [ 8 , 9 , 10 ]. Thereby, the payment providing for the relay nodes makes them be willing to forward the messages for other users.

However, the selfishness of autonomous users may result in the throughput unfairness which only benefits certain users. Tan et al. In their framework, the users who generate higher interference are required to pay more by transmitting at a lower power to give other users a fairer chance of sharing the throughput.

The users could misbehave by broadcasting high price coefficients to force other users to transmit at a lower power without any incentive to play fairly. The traditional networks are built on the assumption that all the network entities cooperate with each other to achieve the desirable network performance or scalability.

However, the assumption may not always found owing to the emergence of some users who change the network behavior in a way in order to benefit themselves at the cost of others.

Sometimes, the node with more ration would only act to achieve an outcome that he gets most. That case is more common in the multihop wireless networks like ad hoc network or sensor network which often consists of wireless battery-powered devices and the networks that need cooperation with each other to complete a task.

Don't show me this again. This is one of over 2, courses on OCW. Explore materials for this course in the pages linked along the left. No enrollment or registration. Freely browse and use OCW materials at your own pace. There's no signup, and no start or end dates. Knowledge is your reward.

Use OCW to guide your own life-long learning, or to teach others. We don't offer credit or certification for using OCW.

These expressiveness results in turn have strong applications in computer science, e. It is easier for me to think of applications of computer science techniques to game theory, than the other way around. There is a very active field of algorithmic game theory which focuses on the development of efficient algorithms or complexity results for, e.

Often, these concepts are easy to define, but prohibitively difficult to compute directly from the definitions. This work extends at least as far as mechanism design, where we attempt to manipulate the rules of auctions in order to guarantee agent behavior e. Noam Nisan, Yoav Shoham, Tim Roughgarden, and many others have some fascinating papers on the subject of mechanism design from a theory point of view; Vince Conitzer has applied AI techniques to the problem to develop automated mechanism design.

On the more applied side in artificial intelligence, it's difficult to think of multi-agent systems without thinking of them as games. The article in Distributed Computing Column 42 attempts to bring a game-theoretic perspective to distributed computing problems.

Game theory and fault tolerance offer two different flavors of robustness to distributed systems — the former is robust against participants attempting to maximize their own utilities, whereas the latter offers robustness against unexpected faults.

This column takes a look at attempts to combine the two. It features a review of recent work that provides both flavors of robustness by Ittai Abraham, Lorenzo Alvisi, and Joe Halpern. Ittai, Lorenzo, and Joe discuss how game theory-style strategic behavior can be accounted for in fault-tolerant distributed protocols.

They make a compelling case for bringing a game-theoretic perspective to distributed computing problems. In Formal Verification game theory is a recurring theme. I think that one of the most important applications is to define the Simulation Preorder as a game between two players: Spoiler he and Duplicator she.

Duplicator has to match the labelled transition and make a move from her starting state. Then, Spoiler makes another move from his last state and Duplicator has to match that transition again, and game goes on in this way.

Spoiler first state simulates Duplicator's first state if she has a winning strategy in this game. In their paper "Advanced automata minimization" , Lorenzo Clemente and Richard Mayr, define a wide variety of simulation relations using games. Since the title is about CS and not TCS, maybe an answer about applications of game theory to networking can be of some interest. Questions about game theory and equilibria arise naturally in networking, since the networks that make Internet are economic competitors and belong to different companies, but they need to collaborate in order to ensure connectivity.

A lot of work was done about routing policy leading to equilibria, and the impact of the price of anarchy on Internet performances. Such problems are generally called "routing games". The seminal work of Roughgarden and Tardos was the starting point of hundreds of papers about equilibria and routing. You can find below some examples mostly works of Eitan Altman or Ariel Orda.

Roughgarden, T. How bad is selfish routing?. Eitan Altman, Jocelyne Elias, Fabio Martignon: A game theoretic framework for joint routing and pricing in networks with elastic demands.

Networking MobiHoc IEEE Trans. In cryptography you often want to get rid of trusted parties.