Modern graph neural networks do worse than classical greedy algorithms in solving combinatorial optimization problems like maximum independent set.
Maria Chiara AngeliniFederico Ricci-TersenghiPublished in: Nat. Mac. Intell. (2023)
Keyphrases
- combinatorial optimization problems
- maximum independent set
- greedy algorithms
- knapsack problem
- combinatorial optimization
- graph theory
- neural network
- discrete optimization
- greedy algorithm
- continuous optimization problems
- graph theoretic
- independent set
- optimization problems
- shortest path problem
- metaheuristic
- traveling salesman problem
- optimal solution
- job shop scheduling
- job shop scheduling problem
- min cost
- exact algorithms
- ant colony optimization
- simulated annealing
- dynamic programming
- branch and bound
- vehicle routing problem
- aco algorithms
- lp relaxation
- branch and bound algorithm
- search algorithm
- genetic algorithm
- minimum spanning tree
- directed acyclic graph
- social network analysis
- graphical models
- np hard
- minmax regret