Irregular mobility: A dynamic alliance formation incentive mechanism under incomplete information

Abstract

The cornerstone of Mobile Crowdsensing is participant mobility, which means different participants will arrive and leave the system separately and their costs also change over time. The uncertainty of participants caused by participants irregular mobility will generate incomplete information, and requesters will be unaware of participants' arrival times, departure times, and cost information. Consequently, the match between requesters and participants by requesters-centric matching algorithm is infeasible because requesters cannot decide on matching strategies because of the lack of knowledge about which participants can be matched. Besides, on account of participant irregular mobility, the match is also unstable due to the irregular changes in participants, requesters may become dissatisfied with the original matching strategies. For incomplete information, a participants-centric matching algorithm where participants have the dominant power is proposed to eliminate the impact of participants' uncertainty. As to the instability of the match, a requesters classification algorithm that could reduce the computational complexity to polynomials is used to update the matching rules. We also promise that with irregular mobility in our algorithm, whichever requester, if not matched in this stage will certainly match participants in the next stage.