Sunil Simon

Choosing products in social networks

We study the consequences of adopting products by agents who form a social network. To this end we use the threshold model introduced in [1], in which the nodes influenced by their neighbours can adopt one out of several alternatives, and associate with each such social network a strategic game between the agents. The possibility of not choosing any product results in two special types of (pure) Nash equilibria. We show that such games may have no Nash equilibrium and that determining the existence of a Nash equilibrium, also of a special type, is NP-complete. The situation changes when the underlying graph of the social network is a DAG, a simple cycle, or has no source nodes. For these three classes we determine the complexity of establishing whether a (special type of) Nash equilibrium exists. We also clarify for these categories of games the status and the complexity of the finite improvement property (FIP). Further, we introduce a new property of the uniform FIP which is satisfied when the underlying graph is a simple cycle, but determining it is co-NP-hard in the general case and also when the underlying graph has no source nodes. The latter complexity results also hold for verifying the property of being a weakly acyclic game.

Nash Equilibrium in Generalised Muller Games

We suggest that extending Muller games with preference ordering for players is a natural way to reason about unbounded duration games. In this context, we look at the standard solution concept of Nash equilibrium for non-zero sum games. We show that Nash equilibria always exists for such generalised Muller games on finite graphs and present a procedure to compute an equilibrium strategy profile. We also give a procedure to compute a subgame perfect equilibrium when it exists in such games.

Reasoning with protocols under imperfect information

We introduce and study a PDL-style logic for reasoning about protocols, or plans, under imperfect information. Our paper touches on a number of issues surrounding the relationship between an agent’s abilities, available choices, and information in an interactive situation. The main question we address is under what circumstances can the agent commit to a protocol or plan, and what can she achieve by doing so?

Social network games

htmlabstractOne of the natural objectives of the field of the social networks is to predict agents’ behaviour. To better understand the spread of various products through a social network [2] introduced a threshold model, in which the nodes influenced by their neighbours can adopt one out of sev- eral alternatives. To analyze the consequences of such product adoption we associate here with each such social network a natural strategic game between the agents. In these games the payoff of each player weakly increases when more players choose his strategy, which is exactly opposite to the congestion games. The possibility of not choosing any product results in two special types of (pure) Nash equilibria. We show that such games may have no Nash equilibrium and that determining an existence of a Nash equilibrium, also of a special type, is NP-complete. This implies the same result for a more general class of games, namely polymatrix games. The situation changes when the underlying graph of the social network is a DAG, a simple cycle, or, more generally, has no source nodes. For these three classes we determine the complexity of an existence of (a special type of) Nash equilibria. We also clarify for these categories of games the status and the com- plexity of the finite best response property (FBRP) and the finite im- provement property (FIP). Further, we introduce a new property of the uniform FIP which is satisfied when the underlying graph is a simple cy- cle, but determining it is co-NP-hard in the general case and also when the underlying graph has no source nodes. The latter complexity results also hold for the property of being a weakly acyclic game. A preliminary version of this paper appeared as [19]

Coordination Games on Graphs (Extended Abstract)

We introduce natural strategic games on graphs, which capture the idea of coordination in a local setting. We show that these games have an exact potential and have strong equilibria when the graph is a pseudoforest. We also exhibit some other classes of graphs for which a strong equilibrium exists. However, in general strong equilibria do not need to exist. Further, we study the (strong) price of stability and anarchy. Finally, we consider the problems of computing strong equilibria and of determining whether a joint strategy is a strong equilibrium.

A Classification of Weakly Acyclic Games

Weakly acyclic games form a natural generalization of the class of games that have the finite improvement property (FIP). In such games one stipulates that from any initial joint strategy some finite improvement path exists. We classify weakly acyclic games using the concept of a scheduler recently introduced in [1].

Coordination Games on Directed Graphs

We study natural strategic games on directed graphs, which capture the idea of coordination in the absence of globally common strategies. We show that these games do not need to have a pure Nash equilibrium and that the problem of determining their existence is NP-complete. The same holds for strong equilibria. We also exhibit some classes of games for which strong equilibria exist and prove that a strong equilibrium can then be found in linear time.

Dynamic restriction of choices: a preliminary logical report

We study games in which the choices available to players are not fixed, and may change during the course of play. Specifically, we consider a model in which players may switch strategies, and a global (social) decision may remove some choices, based on the strategies being adopted by players. We propose a logical formalism in which such choices are specified, and a model of bounded memory strategies in which the eventual implications of such choices can be computed, and present preliminary results.

Paradoxes in social networks with multiple products

We show that various paradoxes can arise in a natural class of social networks. They demonstrate that more services or products may have adverse consequences for all members of the network and conversely that restricting the number of choices may be beneficial for every member of the network. These phenomena have been confirmed by a number of empirical studies. In our analysis we use a simple threshold model of social networks introduced in Apt and Markakis (2011), and more fully in Apt and Markakis (2014). In this model the agents, influenced by their neighbours, can adopt one out of several alternatives. We identify and analyze here four types of paradoxes that can arise in these networks. These paradoxes shed light on possible inefficiencies arising when one modifies the sets of products available to the agents forming a social network or the network structure. One of the paradoxes corresponds to the well-known Braess paradox in congestion games and shows that by adding more choices to a node, the network may end up in a situation that is worse for everybody. We exhibit a dual version of this, according to which removing a product available to an agent can eventually make everybody better off. The other paradoxes that we identify show that by adding or removing a product from the choice set of an agent may lead to permanent instability. Finally, we also identify conditions under which some of these paradoxes cannot arise.

Social Network Games with Obligatory Product Selection

Recently, Apt and Markakis introduced a model for product adoption in social networks with multiple products, where the agents, influenced by their neighbours, can adopt one out of several alternatives (products). To analyze these networks we introduce social network games in which product adoption is obligatory. We show that when the underlying graph is a simple cycle, there is a polynomial time algorithm allowing us to determine whether the game has a Nash equilibrium. In contrast, in the arbitrary case this problem is NP-complete. We also show that the problem of determining whether the game is weakly acyclic is co-NP hard. Using these games we analyze various types of paradoxes that can arise in the considered networks. One of them corresponds to the well-known Braess paradox in congestion games. In particular, we show that social networks exist with the property that by adding an additional product to a specific node, the choices of the nodes will unavoidably evolve in such a way that everybody is strictly worse off.