What is the merit of reinforcement learning for resource allocation problems (IoT)

[dense stream]

when it comes to resource allocation problems, reinforcement learning is always the first ML method that comes to mind. Why though?

[zealous scroll]

Because there is no way to create a sensible set of data for SL. The system needs to try allocation combinations in different environmental conditions and collect the data as to whether it is a good or bad combination at the same time. This is the basic idea behind the RL problem.