Data aggregation techniques exploit spatial and temporal
correlations among data and aggregate data into a smaller
volume as a means to optimize usage of limited network resources
including energy. There is a trade-off among the Quality of
Information (QoI) requirement and energy consumption for computation
and communication. We formulate the energy-efficient
data aggregation problem as a non-linear optimization problem
to optimize the trade-off and control the degree of information
reduction at each node subject to given QoI requirement. Using
the theory of duality optimization, we prove that under a set of
reasonable cost assumptions, the optimal solution can be obtained
despite non-convexity of the problem. Moreover, we propose a
distributed, iterative algorithm that will converge to the optimal
solution. Extensive numerical results are presented to confirm
the validity of the proposed solution approach.