| Many Internet of Things (IoT) applications are able to offer different Quality of Experience (QoE) to users depending on available resources. For example, a low resolution image of a target may be acceptable when IoT devices have insufficient energy. With more resources, IoT devices can acquire and transmit higher resolution images. Such applications are said to support so called {\em imprecise computation}, where their data requirement consists of a {\em mandatory} and an {\em optional} part. As its name implies, mandatory data must be collected whereas optional data is subject to available resources. In this respect, we consider an IoT network operator that uses Unmanned Aerial Vehicles (UAVs) to collect data from one or more geographical regions as per the data requirement of tasks. In this respect, in order to maximize the computational quality of tasks, which is a function of the amount of optional data collected and executed by the operator, this paper outlines a novel Mixed Integer Linear Program (MILP) to compute the data collection schedule of UAVs. Additionally, it introduces BestLoc, a heuristic solution that first assigns UAVs to fulfill the mandatory data requirement of tasks. After that, it assigns UAVs to collect optional data. The results indicate that BestLoc achieved approximately 92.31\% of the optimal computational quality obtained by MILP. |