Inability of a graph neural network heuristic to outperform greedy algorithms in solving combinatorial optimization problems like Max-Cut.
Stefan BoettcherPublished in: CoRR (2022)
Keyphrases
- combinatorial optimization problems
- greedy algorithms
- knapsack problem
- max cut
- combinatorial optimization
- np complete problems
- neural network
- greedy algorithm
- graph model
- graph coloring
- planar graphs
- job shop scheduling
- np hard
- graph partitioning
- optimal solution
- spectral graph
- np complete
- metaheuristic
- dynamic programming
- optimization problems
- job shop scheduling problem
- search heuristics
- greedy heuristic
- lp relaxation
- ant colony optimization
- traveling salesman problem
- weighted graph
- simulated annealing
- search algorithm
- graph structure
- branch and bound algorithm
- combinatorial problems
- graph matching
- vehicle routing problem
- bipartite graph
- branch and bound
- tabu search
- random walk
- search space
- genetic algorithm
- undirected graph
- learning algorithm
- computational complexity
- lower bound
- minimum weight
- min cut
- pairwise
- cost function
- directed graph
- minimum spanning tree
- edge weights
- spanning tree