Joint Task Allocation and Data Delivery Framework for Unmanned Aerial Vehicles in Aerial Plant Inspection

Emerging unmanned aerial vehicle (UAV) or drone technology has brought a great potential to oil and gas industry in which drones can be used for providing aerial plant and facility inspection services. With practical constraints such as a limitation on the number of drones, task allocation is required to address servicing important missions, i.e., delivering aerial data to a con- trol center, with a deadline. The literature on task allocation and data delivery has been studied either deterministic problems, i.e., information is completely known, or separate problems dealing with uncertainties. However, missions to be completed and data to be delivered are dependently uncertain because of emergencies in reality. The uncertainties in task allocation and data delivery need to be considered. Therefore, we formulate a stochastic optimization model to optimize the allocation of drones and aerial data delivery under uncertainties to minimize operating costs. To address a probability of deadline violation, we reformulate the problem as a cardinality formulation. Furthermore, we conduct the performance evaluation by using Solomon datasets and real oil and gas industry sites.