Smart Irrigation using Reinforcement Learning

Problem definition

The aim is to train a reinforcement learning agent to be able to estimate the proper amount of irrigation needed at each time unit in order to keep the moisture of a particular soil layer (the plant root zone let's say) in a certain bound. The agent should use the obtained information from the moisture of different soil layers.