As the gig economy continues to grow, behaviors of workers on gig service platforms have an increasing impact on service satisfaction. For example, fatigue driving behaviors of drivers in ride-hailing platforms may cause serious damages, both for individuals and society. Therefore, regulating behaviors of workers is urgent and challenging. A lot of studies are conducted to detect workers' noncompliance behaviors, such as detecting fatigue driving by computer vision or pattern recognition methods. However, few of them indicate how to efficiently exploit the detection results to regulate workers' behaviors. In this paper, we point out that workers' noncompliance behaviors and their incomes should be correlated, and propose a quantifiable computation framework that includes a price-based incentive mechanism and a method to verify the effectiveness of the mechanism. Historical behaviors of workers are summarized as credits and stored in nonfungible token called CreditToken to ensure that it cannot be tampered with. CreditToken will further affect workers' incomes. We abstract the decision-making behavior of workers as a Markov decision process and demonstrate the effectiveness of the incentive mechanism with model checking and formal methods. The analysis shows that our framework is able to provide a rational price strategy formation for gig service platforms, and can be flexibly integrated into existing pricing schemes to maximize the value of the detection results. Extensive experiments illustrate the advanced nature and practicality of our framework.

中文翻译：

---

监管意识系统的动态激励机制设计

随着零工经济的持续增长，员工在零工服务平台上的行为对服务满意度的影响越来越大。例如，网约车平台司机的疲劳驾驶行为可能对个人和社会造成严重损害。因此，规范工人的行为是紧迫的和具有挑战性的。进行了大量研究以检测工人的不合规行为，例如通过计算机视觉或模式识别方法检测疲劳驾驶。然而，很少有人指出如何有效地利用检测结果来规范工人的行为。在本文中，我们指出工人的不合规行为与其收入应该相关，并提出了一个可量化的计算框架，包括基于价格的激励机制和验证机制有效性的方法。工人的历史行为被汇总为信用并存储在称为 CreditToken 的不可替代的令牌中，以确保它不能被篡改。CreditToken 将进一步影响工人的收入。我们将工人的决策行为抽象为马尔可夫决策过程，并通过模型检查和形式方法证明了激励机制的有效性。分析表明，我们的框架能够为演出服务平台提供合理的价格策略形成，并且可以灵活地集成到现有的定价方案中，以最大化检测结果的价值。