Abstract:

In order to help crowdsourcing platform users search for right tasks conveniently and accurately, promote their ability and solve the dynamic problem of crowdsourcing task recommendation, we propose a crowdsourcing task recommendation system based on multi-agent. Firstly, we build the multi-agent task recommendation system model based on the crowdsourcing platform, and describe the design thought and framework of the model. In addition, we expound the function of each agent, the interaction relations between agents and the related algorithms. Secondly, we propose some related algorithms to promote crowdsourcing users’ ability. Finally, we use the simulation software NetLogo to verify the model. The results show that the proposed crowdsourcing task recommendation system can make contribution to the promotion of users’ ability, which proves the necessity of task recommendation systems on crowdsourcing platforms. Our analysis shows that the multi-agent technique can improve the overall performance of the task recommendation system, such as dynamism, intellectuality and flexibility. It can also promote the management and maintenance of crowdsourcing platform data.

Key words: crowdsourcing design, task recommendation, agent, simulation