A Expectation-Aware Dynamic Pricing Model for Spatial Crowdsourcing Platforms

Spatial crowdsourcing software are becoming indispensable to modern people's lives. Among all topics over spatial crowdsourcing platforms, pricing method study is a fundamental one. The pricing strategy largely influences the incentives of workers and the revenue of platforms. The existing studies only consider the pricing strategy for users, but ignore the pricing to workers. They usually use a fixed percentage multiplied the price by to pay the workers given by users which neglect the status of the workers. However, the expectation price of the workers is also important. On one hand, if the payment to the workers is too high, the platform will reduce the revenue. On the other hand, if the payment to the workers is too low, the workers would reduce their enthusiasm to serve the spatial crowdsourcing tasks, leading to low quantity and quality of completed tasks. In order to balance the platform's revenue and the worker's enthusiasm, we propose a dynamic pricing method based on workers' expectation. Specifically, we consider three factors that may affect the worker's pricing, which are the historical pricing law, the current working time of workers, and the price of the most recent orders. Through extensive experiments on real datasets, we show that our pricing method can improve both the revenue of the platforms and the ratio of completed orders, which means that it can well balance the interests of the platform and the interests of the workers.