Pointer Networks with Q-Learning for Combinatorial Optimization

Alessandro Barro

We introduce the Pointer Q-Network (PQN), a hybrid neural architecture that integrates model-free Q-value policy approximation with Pointer Networks (Ptr-Nets) to enhance the optim…