Reinforcement learning simulation results for the time period 01/02/2021–28/02/2021.

The initial value is the measured RE (left) UER (right) at the start of the period. The difference is the model performances calculated as measured value at the end of the period (i.e. as a result of policymaker) minus the one predicted by implementing the model’s recommended policies (higher is better, where positive values show the policy intervention of the RLA model improved the RE or UER). (XLSX)