Section: New Results

Orchestration of crowdsourcing activities

Participants : François Charoy, Kahina Bessai.

Crowdsourcing is an important paradigm in human problem solving using the Web. When they face a workload outburst, businesses may choose to outsource some or all of their process tasks to the crowd in order to maintain the quality of service promised for their customers. This may occur in situations like crisis management, when organizations are overloaded by a sudden event breakout. These tasks are generally difficult to implement as solution based on software service only. So, the use of crowdsourcing platform seems enticing. To ensure efficient and wise use of resources, methods assisting decision making need to be developed whose aim is to assist businesses in choosing the most knowledgeable workers. We addressed the resource allocation problem in crisis context by defining a delegation approach based on crowdsourcing as resource provider. We introduce a mathematical model for business process execution in crowd-sourcing context and an exact optimization algorithm. As the problem addressed is NP-complete, we proposed a more efficient algorithm that we validated through simulation [7]. Furthermore, to overcome the limitations of existing works we take into account the fact that business process tasks are ordered while optimizing the overall execution time of a given business process instance under budget constraint. We used a synthetic crowd model or valitation. We have also defined a model to validate our work for geo-crowdsourcing activities [8].