# Résolution de problèmes algorithmiques

Find a longest path in a tree.

### Dynamic programming

You can choose some vertex as a root, and then do dynamic programming in each subtree. Let A[v] be the length of the longest path in the subtree rooted at v, and starting at v. In other words A[v] is the depth of the subtree. Let B[v] be the longest path in the subtree, without any restriction. Then A[v]=B[v]=0 when v is a leaf and otherwise

A[v] = max over descendants u of A[u] + 1
B[v] = max of A[v]
A[u1] + 2 + A[u2] where u1≠u2 are descendants
B[u] where u is a descendant


With DFS given a vertex v you can find a furthest vertex v from u. Start with an arbitrary vertex r to find a furthest vertex u. Then start from u to find a furthest vertex v. The u-v path is a longest path. It is a nice exercise to prove this claim.