Information Technology Reference
In-Depth Information
Chapter 2
Social Group Utility Maximization Framework
2.1
Motivation
As discussed in Chap. 1, the social ties among wireless users significantly influence
their interactions with each other in wireless networks. One fundamental aspect of
the positive social tie between two users is that they are altruistic to each other
such that one cares about the other's welfare. As a result, a user would take into
account the effect of its behavior on those having social ties
1
with it. It is then
natural to ask “Is it possible to exploit users' social ties to stimulate their cooperative
behaviors?” Indeed, altruistic behaviors are often observed among people with social
ties. With this motivation, we view a wireless network as an overlay/underlay system
(as illustrated in Fig.
2.1
), where a “virtual social network” (social domain) overlays
a physical communication network (physical domain). Wireless users are connected
by social ties in the social domain, while their wireless devices are subject to physical
relationships in the physical domain. It is important to observe that users generally
have
diverse
social ties such that a user cares about others at different levels. For
example, a user may care about her family members more than her friends, and
cares about her friends more than an acquaintance of her. Similarly, it is clear that
wireless devices also generally have
diverse
physical relationships. For example,
depending on their physical locations, wireless devices can cause different levels of
interference to each other. A primary goal here is to leverage the intrinsic diverse
social tie structure among wireless users, which can be viewed as “hidden incentives”
based on existing human relationships, to facilitate cooperative networking among
their wireless devices subject to diverse physical relationships.
To this end, we advocate a social group utility maximization (SGUM) framework
that takes into account both the diverse social coupling and diverse physical coupling
among users. Specifically, we model the social coupling and physical coupling by
a social graph and physical graph, respectively, and then we cast the distributed
decision making problem among users as a SGUM game.
