TJ1812 / Adaptive-Traffic-Signal-Control-Using-Reinforcement-Learning

This is an application exploiting principles of Deep Reinforcement Learning. The Deep Neural Network is trained to approximate the Bellman Equation (Q-Learning).
133 stars 47 forks source link

How to compare this model ? #5

Open junaidmumin opened 5 years ago

junaidmumin commented 5 years ago

Sir as for as the report is concerned you have analyzed that the best result we got was around 90k. It is about 72% improvement. The average waiting times are about 160-200k which is still a significant improvement from our traditional methods. My question to you is how you have analyzed as I am not able to get these statistics and don't know how to get this total time? Can you please send the detailed work