Coverage and Workload Cost Balancing in Spatial Crowdsourcing

This paper addresses the coverage and workloadbalancing requirements of worker recruiting in spatial crowdsourcing. That is, the recruited workers should be able to visit all the crowdsourcing locations to satisfy a certain quality, e.g., traffic monitoring or climate forecast. In addition, each crowdsourcing operation has a cost, e.g., data traffic or energy consumption, and each crowdsourcing location might have a crowdsourcing budget for the visited workers. The objective of this paper is to find a worker recruiting algorithm, which ensures the coverage requirement and minimizes the maximal crowdsourcing cost for any crowdsourcing location. We gradually discuss the problem from the 1-D scenario to the general 2-D scenario. In the 1-D scenario, we propose a bounded directional greedy algorithm first. Then, we propose a PTAS extension. A dynamic programming solution is further proposed with a higher computation complexity. In the 2-D scenario, we propose a randomized rounding algorithm with an O( logn log logn ) approximation ratio in a high probability. Extensive experiments on realistic traces demonstrate the effectiveness of the proposed algorithms.