A double-layer crowd evacuation simulation method based on deep reinforcement learning.
Yong ZhangBo YangJianlin ZhuPublished in: Comput. Animat. Virtual Worlds (2024)
Keyphrases
- reinforcement learning
- dynamic programming
- cost function
- computational cost
- synthetic data
- detection method
- high accuracy
- support vector machine
- experimental evaluation
- preprocessing
- multi agent
- classification method
- function approximation
- real robot
- mathematical model
- cellular automata
- neural network
- emergency evacuation
- clustering method
- input data
- significant improvement
- prior knowledge
- computational complexity
- multiscale