Help with the critical demands

I have one question, I have been training my agent with the descriptions in the paper, but I had a question in my head. What would happend if I increase the percentage of critical demands?

In this case I noticed that the agent doesn't train like with the 15%. Why is that number so important in the training or in the paper that I can't change it. Or, in other case, what I should do or what other hiperparameters should I also change in order to have another good agent?

Thank you.

The issue asks about hyperparameter tuning for the 'critical demands' percentage in the ENERO reinforcement learning agent. Start by reviewing the paper and code to understand the role of this parameter in the PPO algorithm and network routing optimization. Examine the training scripts and configuration files to see how the 15% value is set and what other hyperparameters might interact with it. A good answer would involve running experiments with different percentages and observing the agent's training performance.

Written by the indexing model from the issue text.