In the past decade the have been many developments in the field of Internet of Things (IoT), including its implementations as smart houses, smart transportation, smart cities etc. In these implementations, there are more and more devices connected to the internet, but they usually have limited resources. Several papers dealt with graph-implantations of IoT object and web services, and different graph algorithms applied on these structures, such as the shortest path problem, or finding all paths from source point to target point. This research describes a more complex graph, in which the nodes (objects) are knapsack items, thus having two additional independent (from the edge weights-that represent physical distance) attributes, one is limited by a certain constant (the knapsack limit-which is the data size), and the other is optimally minimized (processing time). The decision making has new aspects in this kind of graph, and the choice of objects in an optimal manner in this graph is the main subject of this paper.