Worker Selection for Reliably Crowdsourcing Location-Dependent Tasks

Obtaining accurate information about specific locations is of great importance to today's many crowdsourced smartphone applications. To verify information about a location, smartphone users are selected to go to the location and answer a yes/no question about the location. Our research focuses on the location-aware worker selection problem, which is the problem of selecting a group of workers who, together, can give the most accurate answer to the location-based question. We define the location-aware worker selection problem, mathematically formulate it, and then show that an optimal solution is exponential in time complexity. We present our heuristic solutions that take into account both the reliability of the users and the level of convenience for each user to complete the task. We evaluate and compare our approaches to three other heuristic algorithms via simulation.