the mechanism is Dominant Strategy Incentive Compatible (DSIC) or strategy-proof.
One way used in mechanism design to make the agents to reveal true private information
is to incentivise them with payments. However, the payments causes loss in overall
utility of agents. Therefore, in applications where there is no owner of the resource who
will collect payments as revenue, and the users or systems involved wanted to efficiently
allocate resource among themselves, the payments are undesirable.
A mechanism is
budget balanced (BB) if the net payment from agents to social planner or budget surplus
is zero. Thus, in situations considered here, like the resource allocation problem in the
previous section, ideally we would like to efficiently allocate the resource to users using
true values reported by them while making the loss in utility due to payments zero.
1.3
Budget balancing problem
Groves mechanisms [8] are the only allocatively efficient and strategy-proof mechanisms
in quasi linear environment. However, they are not budget balanced. In fact, the Vickrey-
Clarke-Groves (VCG) mechanism (see [9] and [10]), a member of Groves class of mecha-
nisms, maximizes the total payments from the agents to the social planner. The Green-
Laffont impossibility theorem [11] says that there is no mechanism in a quasi-linear
environment that is strategy-proof, achieves allocative efficiency, and is budget balanced.
An incentive compatible mechanism gives appropriate incentive to users for eliciting
truth from them. These mechanisms require a commodity for which agents have ex-
tremely large demand known as 'numeraire commodity' to incentivise them. Agents are
charged in terms of this numeraire commodity for their consumption of original resource
to be allocated, such that their overall utility is maximized when they report true infor-
mation. In pricing schemes, a common way is to model money as a numeraire commodity
4