Research On The Optimal Allocation Of Crowdsourcing Task And The Reputation Model Of Crowdsourcing Workers

In the process of crowdsourcing task allocation,the two most important factors are crowdsourcing workers and crowdsourcing tasks.Every crowdsourcing worker has different environmental backgrounds,knowledge skills and attributes,each crowdsourcing task has the demand for different attributes of professional skills.Under the environment of new mobile Internet,crowdsourcing workers' s abilities are different,with nonfeterminacy and multiformity.Designing crowdsourcing tasks reasonably and allocating appropriate crowdsourced workers are of great significance to improve the quality of results task,save the time cost and economic cost of crowdsourcing companies.Traditional task allocation strategy is "one to one",which means one work is only allocated to one worker.Crowdsourcing task allocation strategy is adhering to the “many to many”allocation principles.In addition,the traditional crowdsourcing task allocation model ignores the task charateristic as well as worker,making the lack of connection between task properties charateristic and worker attributes.Crowdsourcing optimized allocation model should not only consider the relationship between crowdsourcing tasks,the association between task properties charcteristic and crowdsourcing worker attributes should also be considered.In addition,the traditional crowdsourcing platform is just hiring the workers,and lacks of mechanism to guide and motivate workers to complete the task efficiently,can result in the win-win pattern between workers and the publisher of the tasks,and is an important factor to improve the performance of crowdsourcing system.The design and allocation model of crowdsourcing tasks on the collection platform of geographic interest point informations were studied in this thesis,and provide the design and allocation model of crowdsourcing task into packages,the crowdsourced workers reputation model based on active degree and the crowdsourcing optimiazetion allocation model based on both sides' s satisfaction.The works done in this thesis are as follows:1)Proposed the design model to cluster the crowdsourcing task into packages according to the geographical position of these tasks.which is aimed at the shortcomings of design model of crowdsourcing tasks on the acquisition platform of geographic interest point informations.Also designed remuneration incentive mechanism of crowdsourcing packages and summarized the attributes of crowdsourcing packages.2)Proposed the crowdsourcing workers reputation model based on active degree and analyzed the parameter value of this model in the background of the design of crowdsourcing packages on the collection platform of geographic interest point informations,which is aimed at the shortcomings of the classical worker average reputation model.3)Proposed the crowdsourcing task optimization allocation model based on the satisfaction of crowdsourcing workers and packages,which is aimed at the shortcomings of traditional crowdsourcing task allocation mechanism in acquisition platform of geographical interest point information,and give full consideration to the correlation between the attributes of crowdsourcing workers and task packages.4)Designed contrast experiments to verify the feasibility of the design strategy which packs tasks into task packages,the effectiveness of the workers reputation model based on active degree and the optimization allocation model based on the satisfaction of crowdsourcing workers and packages on the acquisition platform of geographical interest point information.The experimental tests show that the design strategy which packs tasks into task packages,the workers reputation model based on active degree and the optimization allocation model based on the satisfaction of crowdsourcing workers and packages provide in this thesis can balance the distribution of update cycle under different density of interest points,well organize and motivate workers to do the crowdsourcing tasks,get high complete quality of crowdsourcing task.And they also have a certain practical value and reference significance to the use of crowdsourcing in acquiring geographical point of interest information.