PUWR-TSSG: A CMAB-based post-unknown worker recruitment scheme for Three-Stage Stackelberg Games in Mobile Crowd Sensing

Abstract

Numerous Three-Stage Stackelberg Games (TSSG) have been proposed to model the strategic interactions among the requesters, the platform, and the workers in Mobile Crowd Sensing (MCS). However, most existing studies unrealistically assume that the platform possesses prior knowledge of the workers’ credibility either beforehand or after receiving their data. Conversely, in practical scenarios, the credibility of workers remains uncertain even after the submission of their data, which is known as the Post-Unknown Worker Recruitment (PUWR) problem. Given this context, conventional models designed for TSSG cannot be applied to real-world MCS. In this paper, we present the PUWR-TSSG scheme for quality-enhanced worker recruitment in TSSG. Specifically, we avoid the unreasonable assumption in previous works and propose a Double-level Credibility Discovery (DCD) approach with bipartite graph-based matrix completion for accurate credibility verification. Subsequently, based on the DCD approach, we further propose a meticulously designed combinatorial multi-armed bandit mechanism to solve the exploration–exploitation dilemma in untrusted environments. Furthermore, we formulate the payment computation issue as a TSSG, while simultaneously considering the workers’ credibility and verification costs incurred by the PUWR problem. Theoretical analyses validate the existence of Stackelberg Equilibrium in our scheme, ensuring that no participant has an incentive to unilaterally deviate from its optimal strategy. Extensive simulations on a real-world dataset validate the effectiveness of our proposed PUWR-TSSG scheme, significantly enhancing the overall data quality and leading to a remarkable average reduction in regret of up to 85.9% compared to baseline methods.