Research On Collective Allocation Method Of Crowdsourcing Complex Tasks And Worker Strategy Synchronization Mechanism

Crowdsourcing is a new type of task assignment model,which utilizes the intelligence distribution and collaboration of Internet mass groups to complete tasks.With the increasing complexity of tasks in crowdsourcing platforms,how to find better allocation methods and explore the cooperation mechanism of workers when executing tasks has become a hot issue in the research of complex crowdsourcing task allocation,and bringing new challenges to crowdsourcing research.This thesis takes the complex task allocation and worker cooperation mechanism as the research goal,conducts in-depth research on the global allocation of large-scale complex tasks,and explores the synchronization and coordination of workers in the cooperative execution of tasks and the mechanism of the prominent phenomenon of workers.Thus,this thesis proposes a collective assignment model for crowdsourcing tasks and a strategy coordination synchronization model for workers.The main contributions and innovations of this thesis are summarized in the following two aspects:Firstly,this thesis establishes a collective allocation model for complex crowdsourcing tasks based on bipartite graph matching.Previous research work only focused on how to find a worker team that meets the task requirements,without comprehensively considering factors such as task requirements、worker skills、time、budget and reward of worker.When assigning large-scale complex tasks,relatively complex tasks will not be able to find workers who meet the requirements,resulting in allocation failures.In order to make up for the limitations of related research,The model proposed in this paper can solve the problem of low success rate of concurrent assignment of large-scale and complex tasks in the platform,and can make reasonable use of workers in the platform.In the model,the complex task allocation problem is transformed into a weighted bipartite graph matching problem,and then the optimal allocation scheme is solved by using the KM(Kuhn-Munkres)algorithm.In this way,as many tasks as possible can be assigned to workers.This thesis conducts comparative experiments on real data sets,and the results show that the proposed model has better performance in terms of task success rate and total task payment.Secondly,this thesis extends the Multi-agent system group behavior to the crowdsourcing environment,and proposes a strategy synchronization coordination model for crowdsourcing workers.In real crowdsourcing,workers are often not independent in executing tasks,and their strategies for executing tasks will be affected by their own preferences,social relations,etc.Most current research only focuses on how to encourage collaboration and communication among workers.The model proposed in this thesis is to study the specific process of workers’ interactive communication and the phenomenon of workers’ prominence decision.In the synchronous coordination model,workers can adaptively adjust and coordinate their own strategies,and we find that strategy of the worker will gather to decision of prominence workers.In this thesis,the model is tested in different environments,and the experimental results verify the correctness of the model proposed in this thesis.