Using Reinforcement Learning for Per-Instance Algorithm Configuration on the TSP.
Moritz Vinzent SeilerJeroen RookJonathan HeinsOliver Ludger PreußJakob BossekHeike TrautmannPublished in: SSCI (2023)
Keyphrases
- detection algorithm
- computational complexity
- high accuracy
- simulated annealing
- optimal solution
- experimental evaluation
- np hard
- dynamic programming
- reinforcement learning
- optimization algorithm
- k means
- particle swarm optimization
- significant improvement
- computational cost
- computationally efficient
- segmentation algorithm
- ant colony optimization
- times faster
- theoretical analysis
- input data
- matching algorithm
- path planning
- combinatorial optimization
- convergence rate
- preprocessing
- recognition algorithm
- objective function
- learning algorithm
- model free
- monte carlo
- expectation maximization
- optimization problems
- worst case
- mobile robot
- multi objective
- search space
- multi agent
- similarity measure