Abstract:The past decade has witnessed rapid development of wireless communication and ubiquitous availability of mobile devices. Spatial crowdsourcing (SC), which requires workers to physically move to specific locations to perform tasks, has gained unprecedented popularity. A typical class of SC tasks is macro tasks, which demand the cooperation of multiple workers with specific skills. Existing studies on macro task assignment often generate an assignment strategy where worker skill set can fully cover the required skills, leading to ambiguous division of labor among workers and unreasonable static distribution of rewards. This work studies the explicit assignment and dynamic pricing (EADP) problem of macro tasks, and considers both online explicit task assignment (skill used and realtime workload) and dynamic reward distribution. Aiming to maximize social welfare of the SC system, we develop a pricing mechanism based on primal-dual technique, and propose an online batch-based framework coupled with an effective dual subroutine to assign macro SC tasks. Through theoretical analysis, we prove that the pricing mechanism is truthful and individually rational, and our online framework guarantees a good parameterized competitive ratio, as well as polynomial time complexity. Extensive experiments demonstrate our algorithms’ effectiveness and efficiency.

Explicit Assignment and Dynamic Pricing of Macro Online Tasks in Spatial Crowdsourcing