-
Nodes: 0

Graham Scan finds the convex hull of a set of nodes. It is a backtracking algorithm which works by ordering nodes by their relative polar angle with the start node.

Kruskal's algorithm is a greedy algorithm for finding a Minimum Spanning Tree. It sorts edges by weight then repeatedly adds the next smallest edge which doesn't create a cycle.

Prim's algorithm is also a greedy algorithm for finding a Minimum Spanning Tree. It starts at a random node then repeatedly adds the edge with the smallest weight which connects an unvisited node to a visited node.

Greedy Minimum Matching finds a minimal matching that connects all pairs of nodes with the minimum total edge weight.

Nearest Neighbour is a greedy algorithm for finding a sub-optimal hamiltonian tour of a graph by repeatedly choosing the nearest unvisited node.

2-Opt performs a local search to improve a sub-optimal tour by reordering a route which crosses over itself.