Redefining Node Centrality for Task Allocation in Mobile CrowdSensing Platforms

Abstract

With the recent developments in Mobile CrowdSensing, an interesting model of temporal graphs has emerged, in which node weights evolve over time, according to the availability of spatio-temporal tasks on the mobility field. The analysis and understanding of these types of graphs, namely Weight Evolving Temporal (WET) graphs, is critical for optimizing task allocation in such crowdsensing platforms. In this paper, we formally define WET graphs and their corresponding routing problem, in which the objective of the routing is to maximize the reward collected from vertices visited amid the graph traversal. By modeling a WET graph as a time-ordered graph, we define efficient and optimal routing algorithms, and theoretically analyze them. Moreover, we present a novel node centrality measure, namely Coverage Centrality, that captures the popularity of various nodes of the WET graph, and which we incorporate in an online crowdsensing task allocation mechanism to increase task coverage. Finally, we evaluate the efficacy of this novel centrality measure on different types of graphs, when compared to other centrality measures, and evaluate its effect on task coverage in online mobile crowdsensing platforms.