Reinforcement Learning for Combinatorial Optimization

Reinforcement Learning for Combinatorial Optimization

1. Encoder의 변화에 따라

Pointer Network

• Neural Combinatorial Optimization with Reinforcement Learning

  Self-attention

• Learning Heuristics for the TSP by Policy Gradient
• Attention, Learn to Solve Routing Problems!
• Dynamic states during decoding
  • Reinforcement Learning for Solving the Vehicle Routing Problem

    Graph Neural Network

• off-policy
  • Learning Combinatorial Optimization Algorithms over Graphs
• Explortation
  • Exploratory Combinatorial Optimization with Reinforcement Learning
• MCTS
  • Learning to Solve Combinatorial Optimization Problems on Real-World Graphs in Linear Time
  • Solving NP-hard Problems on Graphs with Extended AlphaGo Zero