关键词: 空间众包, 周期性任务, 博弈论, 总效用, 熟悉度

Abstract: For a type of repetitive, fixed-demand periodic tasks that exist in spatial crowdsourcing, but the existing matching algorithm ignores the issue of familiarity required for periodic tasks, an online matching algorithm for supporting periodic tasks in spatial crowdsourcing was proposed. First, the online matching problem was regarded as a multiplayer game, with tasks considered as independent participants in the game. The utility functions were determined based on the tasks' preference for matching workers with high familiarity and the workers' preference for tasks with high rewards and short distances, which were then analyzed using game theory. Then, a simulated annealing strategy was introduced into the updating strategy of the game theory, resulting in the design of the Game Theory (GT) based on the simulated annealing strategy. Finally, a matching with greater total utility was achieved upon reaching a Nash equilibrium. Experimental results on a real dataset demonstrate that, compared to the existing optimal Deferred Acceptance Bidirectional Selection (DABS), the total utility of the matching by GT has increased by 14%. GT achieves matching results that better meet the individual needs of periodic tasks and workers, which can enhance user satisfaction on online spatial crowdsourcing platforms.

Key words: spatial crowdsourcing, periodic task, Game Theory (GT), total utility, familiarity