A novel online incentive mechanism under budget constraint for crowdsourcing systems

Abstract

Crowdsourcing is a new paradigm which leverages human computation to finish specific tasks. Efficient incentive mechanisms are proposed to motivate crowd workers' participation, in which however, researchers frequently concentrate on offline forms. Considering worker's successive arrival and dynamic departure, we are motivated to design an online incentive mechanism. Particularly, the workers come one after another to the crowdsourcing platform and meanwhile the platform makes irrevocable decision to select the qualified workers to the winner set and give them monetary reward. We formulate the incentive mechanism as a knapsack secretary auction, where we take advantages of the principles of solving knapsack problem and secretary problem to tackle the budget constraint and online scenario in our case. Some prevalent-used economic properties, such as individual rationality, truthfulness and fairness are theoretically guaranteed. Extensive evaluations are conducted to demonstrate the effectiveness and correctness of our mechanism with respect to the aforementioned properties.