Computing a profit-maximizing sequence of offers to agents in a social network

Firms have ever-increasing amounts of information about possible customers available to them; furthermore, they are increasingly able to push offers to them rather than having to passively wait for a consumer to initiate contact. This opens up enormous new opportunities for intelligent marketing. In this paper, we consider the limit case in which the firm can predict consumers' preferences and relationships to each other perfectly, and has perfect control over when it makes offers to consumers. We focus on how to optimally introduce a new product into a social network of agents, when that product has significant externalities. We propose a general model to capture this problem, and prove that there is no polynomial-time approximation unless P=NP. However, in the special case where agents' relationships are symmetric and externalities are positive, we show that the problem can be solved in polynomial time. © 2012 Springer-Verlag.

Duke Scholars

Author Vincent Conitzer Computer Science