Research On Task Publishing Approach Of Crowdsourcing Workflows

Crowdsourcing represents the act of a company or institution taking a function once performed by employees and outsourcing it to an undefined(and generally large)network of people in the form of an open call.As the crowdsourcing task becomes more and more complex,the current crowdsourcing work mode is to combine crowdsourcing with workflow,and get the solution to the original crowdsourcing work through the following work process of"crowdsourcing work disassembling? workers handling simple sub-tasks? task solutions combination".When publishing the tasks in the crowdsourcing on the platform,the settings of task attributes will affect the overall completion effect of the crowdsourcing work.Therefore,when publishing tasks in crowdsourcing workflows,it is necessary to set a reasonable value of each task attribute,which not only improves the completion efficiency of the workflow,but also reduces the completion cost.This is Crowdsourcing Task Publishing(CTP)problem.At present,there are few studies on CTP problem.In the existing research,the CTP problem is stipulated as a constraint solving problem to find the global optimal value that should be set for each task attribute of the tasks.However,when the number of tasks to be processed is large,the number of constraints will also increase,the solution time will grow rapidly.In a crowdsourcing environment,after the requesters post the task to the crowdsourcing platform,it is often desirable that the task be processed as soon as possible and the long solution time is unacceptable.If the optimization process takes too long,it will affect the overall work progress and reduce work efficiency.Therefore,further discussion on the solution to the CTP problem is needed.The contributions of this paper are as follows:(1)Modeling the framework to solve the CTP problem in workflow,and the main functions of the data analysis module,the optimization module and the task publishing module in the problem solving framework is clarified.The objectives and many constraints of CTP problem that should be considered are determined in consideration of multiple factors that affect the quality of the completion of the crowdsourcing tasks.(2)The optimization module in the CTP problem solving framework needs to optimize the task attributes setting problem.The optimization problem can be stipulated as a quadratic programming problem,which can be optimized using the constraint solving method(defined as CTP-C optimization method in this paper);In order to alleviate the problem that the solution time will increases rapidly with the increase of the problem size in CTP-C optimization method,an optimization method based on greedy algorithm including four heuristic strategies is proposed to optimize the CTP problem(defined as CTP-H in this paper).(3)Through a series of experiments,the effects of different constraint solving tools on the CTP-C optimization method are compared.The reasonable setting of the task publishing condition in the task publishing module is verified through experiment.The feasibility and applicability of the CTP-C optimization method and the CTP-H optimization method were compared and analyzed through experiments.