Closed theomarzaki closed 5 years ago
All applicable to Models:
Exploration vs exploitation = epsilon greedy with 40 iteration for exploration 10 iterations explorations 👍
Exponential reward to Agent (negative vs positive) 👍
Random Forest Feature space reduced to evolve around positions, speed, accuracy excluding spacing for merging (taken care of in reward allocation) 👍
Simulation reflects status of the road with additional parameters (lstm,ddqn,ddqn) 👍
All applicable to Models:
Exploration vs exploitation = epsilon greedy with 40 iteration for exploration 10 iterations explorations 👍
Exponential reward to Agent (negative vs positive) 👍
Random Forest Feature space reduced to evolve around positions, speed, accuracy excluding spacing for merging (taken care of in reward allocation) 👍
Simulation reflects status of the road with additional parameters (lstm,ddqn,ddqn) 👍