Multi-Agent Deep Reinforcement Learning-Based Algorithm For Fast Generalization On Routing Problems

We propose a fast generalization method for DQN-Routing, an algorithm based on multi-agent deep reinforcement learning that suffers from problem when introduced to new topologies even if it was trained similar topology. The proposed is the wisdom of crowds and allows distributed routing algorithm, generalize better were not seen before during training. also aims decrease solution search time as original DQN-Routing takes long converge, increase overall performance by minimizing mean delivery total power consumption number collisions. experimental evaluation our proved capable outperform algorithm.