Towards Mining of Stakeholders in Criminal Organizations from Telecommunication Metadata : Analytic Approach to Latent Feature Extraction

Abstract

High power to guarantee strict performance requirement and low power to avoid energy depletion result in an inevitable conflict for a renewable energy harvesting communication system with finite energy storage. This paper proposes a generic approach to study the per-flow performance in such a multichannel system multiplexed by multiple flows. The queueing delay constraint and energy storage constraint are constructed to express the probabilistic bound of queuing delay and that of energy depletion, respectively. We study these constraints with the statistical information of the processes, including traffic arrivals and service, energy harvesting, and consumption. The lower bound of the long-term maximum per-flow throughput is then derived to meet the constraints under a specific system. The accuracy of the proposed approach is validated by simulation experiments. The analysis reveals how the sustained throughput is affected by various factors, such as the queueing delay and energy storage constraints, the packet size, the energy block size, the traffic scheduling schemes, the bandwidth allocation schemes as well as the interdependence among the channel service processes. Particularly, the analysis also provides valuable insight into traffic admission control from the viewpoint of small queueing delay and finite energy storage.