Research On Task Allocation Algorithm To Cooperative Sensing Based On Auction Theory

Mobile Crowd Sensing System(MCS)is a new paradigm of Internet of Things(Io T)sensing,and its application has become more and more popular.The development of the Io T has spawned a variety of sensing devices with different functions.These sensing devices can sense and upload data of various objects at any time and anywhere,resulting in and promoting the MCS of different dimensions and different data domains.Task allocation needs to coordinate and encourage users to participate in sensing tasks and directly affect the performance of data quality,cost,privacy protection,system efficiency and energy consumption,which is a vital part of MCS.Current research often assumes that users’ covering and sensing abilities are isomorphic,which is inconsistent with the increasing trend of heterogeneous devices in the Io T.At the same time,privacy protection is also a hot issue in MCS.The improvement of users’ privacy security can encourage more users to participate in sensing and expand application scenarios.Based on the above problems,this paper studies the cooperative sensing task allocation algorithm based on auction theory,the specific contents are as follows :Firstly,in the task allocation problem of location covering related to MCS,in order to fill the gap in the heterogeneity of users’ covering ability,a cooperative sensing task allocation algorithm TAGA based on combinatorial auction and genetic algorithm is proposed.This algorithm first conducts pricing and selection of candidate user groups based on VCG combinatorial auction pricing model,and then realizes task allocation through improved Elite-preserving Genetic Algorithm(EGA).Theoretical analysis and experimental simulation show that the proposed algorithm can achieve Bayesian Incentive Compatibility(BIC),Individual Rationality(IR)and Balance Budget(BB)in economics.At the same time,through the comparison and simulation with MTMW,RTMW and MTRW algorithms,the TAGA algorithm proposed in this paper converges faster when it approaches the maximum system utility and the maximum number of tasks assigned,and has a significant improvement in the system utility and the number of tasks completed.Secondly,in order to ensure the quality of task completion and user privacy security,a task allocation algorithm IPTA based on k-anonymous privacy protection is designed for the user collaborative sensing scenario with heterogeneous covering ability and sensing ability.Since the use of third-party institutions to ensure privacy security will bring some information loss,a metamorphic center location aggregation method is used to reduce the minimum information loss.The simulation is carried out on two real datasets,and the experimental results on Gowalla dataset show that the proposed algorithm can significantly reduce the information loss compared with MDAV algorithm under the same k value.The comparative test on the CRAWDAD dataset verifies that the algorithm also has certain improvement in spatial stability,and has higher buyer satisfaction than the KASD algorithm.