Robust and Efficient Mechanism Design for Heterogeneous Task Crowdsensing

Crowdsensing is a novel concept that divides tasks between participants in order to get an accumulated result. To make the crowdsensing system work well and get better quality, it is indispensable to set up incentive mechanisms to get more workers involved. As far as the bidding of heterogeneous combinations is concerned, combinatorial auctions are the natural choice for workers to bid. Truthfulness and efficiency can be guaranteed based on the properties of VCG mechanism which will result in the higher bid price and high overpayment. To overcome this potential shortcoming of the VCG mechanism, we propose the core-selecting mechanism for the heterogeneous task auction under the crowdsensing market. Two payment rules are applied to the core-selecting auction based on linear programming and quadratic programming techniques to minimize the bidders’ incentives which deviate from their truthful-telling. After extensive simulation experiments, it is proved that our model can decrease the cost significantly.