Loyalty-based Task Assignment in Spatial Crowdsourcing

Abstract

With the fast-paced development of mobile networks and the widespread usage of mobile devices, Spatial Crowdsourcing (SC) has drawn increasing attention in recent years. SC has the potential for collecting information for a broad range of applications such as on-demand local delivery and on-demand transportation. One of the critical issues in SC is task assignment that allocates location-based tasks (e.g., delivering food and packages) to appropriate moving workers (i.e., intelligent device carriers). In this paper, we study a loyalty-based task assignment problem, which aims to maximize the overall rewards of workers while considering worker loyalty. We propose a two-phase framework to solve the problem, including a worker loyalty prediction and a task assignment phase. In the first phase, we use a model based on an efficient time series prediction method called Prophet and an Entropy Weighting method to extract workers' short-term and long-term loyalty and then predict workers' current loyalty scores. In the task assignment phase, we design a Kuhn-Munkras-based algorithm that achieves the optimal task assignment and an efficient Degree-Reduction-based algorithm with minority first scheme. Extensive experiments offer insight into the effectiveness and efficiency of the proposed solutions.