In computer science, there is a method called "Divide And Conquer By Node" to solve some hard problems about paths on a tree. Let's desribe how this method works by function:
solve(t) (t is a tree):
This ends when t has only one node because after deleting it, there's nothing.
Now, WJMZBMR has mistakenly believed that it's ok to chose any node in "Line A". So he'll chose a node at random. To make the situation worse, he thinks a "tree" should have the same number of edges and nodes! So this procedure becomes like that.
Let's define the variable totalCost. Initially the value of totalCost equal to 0. So, solve(t) (now t is a graph):
He'll apply solve on a connected graph with n nodes and n edges. He thinks it will work quickly, but it's very slow. So he wants to know the expectation of totalCost of this procedure. Can you help him?
The first line contains an integer n (3 ≤ n ≤ 3000) — the number of nodes and edges in the graph. Each of the next n lines contains two space-separated integers a_{i}, b_{i} (0 ≤ a_{i}, b_{i} ≤ n - 1) indicating an edge between nodes a_{i} and b_{i}.
Consider that the graph nodes are numbered from 0 to (n - 1). It's guaranteed that there are no self-loops, no multiple edges in that graph. It's guaranteed that the graph is connected.
Print a single real number — the expectation of totalCost. Your answer will be considered correct if its absolute or relative error does not exceed 10^{ - 6}.
5
3 4
2 3
2 4
0 4
1 2
13.166666666666666
3
0 1
1 2
0 2
6.000000000000000
5
0 1
1 2
2 0
3 0
4 1
13.166666666666666
Consider the second example. No matter what we choose first, the totalCost will always be 3 + 2 + 1 = 6.
Name |
---|