Complex Task Design Based on Crowdsourcing Using Petri Net

Complex Task Design Based on Crowdsourcing Using Petri Net

Design of complex crowdsourcing tasks is a challenging problem. This study presented an alternative method called CRVA to address the challenge. The CRVA used hierarchical Petri nets to model the issue. The requester first constructed a root net, and skilled workers then refined composite tasks in the root net by sub-nets. The final outputs were composed of a root net and one or more sub-nets. When a composite task was activated, the control turned to the corresponding sub-net; when the sub-net completed its operation, the control passed back to the parent net. The new method improves task design by enabling ordinary requesters to delegate complex design tasks to specialized crowds, and automatically checks the output quality based on Petri nets with precise formal semantics. Results showed the proposed method is not only feasible but also advantageous compared to other decomposition-based methods.