You are given an integer n, representing n nodes numbered from 0 to n - 1 and a list of edges, where edges[i] = [ui, vi, si, musti]:
ui and vi indicates an undirected edge between nodes ui and vi.
si is the strength of the edge.
musti is an integer (0 or 1). If musti == 1, the edge must be included in thespanning tree. These edges cannot be upgraded.
You are also given an integer k, the maximum number of upgrades you can perform. Each upgrade doubles the strength of an edge, and each eligible edge (with musti == 0) can be upgraded at most once.
The stability of a spanning tree is defined as the minimum strength score among all edges included in it.
Return the maximum possible stability of any valid spanning tree. If it is impossible to connect all nodes, return -1.
Note: A spanning tree of a graph with n nodes is a subset of the edges that connects all nodes together (i.e. the graph is connected) without forming any cycles, and uses exactlyn - 1 edges.
Β
Example 1:
Input:n = 3, edges = [[0,1,2,1],[1,2,3,0]], k = 1
Output:2
Explanation:
Edge [0,1] with strength = 2 must be included in the spanning tree.
Edge [1,2] is optional and can be upgraded from 3 to 6 using one upgrade.
The resulting spanning tree includes these two edges with strengths 2 and 6.
The minimum strength in the spanning tree is 2, which is the maximum possible stability.
Example 2:
Input:n = 3, edges = [[0,1,4,0],[1,2,3,0],[0,2,1,0]], k = 2
Output:6
Explanation:
Since all edges are optional and up to k = 2 upgrades are allowed.
Upgrade edges [0,1] from 4 to 8 and [1,2] from 3 to 6.
The resulting spanning tree includes these two edges with strengths 8 and 6.
The minimum strength in the tree is 6, which is the maximum possible stability.
Example 3:
Input:n = 3, edges = [[0,1,1,1],[1,2,1,1],[2,0,1,1]], k = 0
Output:-1
Explanation:
All edges are mandatory and form a cycle, which violates the spanning tree property of acyclicity. Thus, the answer is -1.
Β
Constraints:
2 <= n <= 105
1 <= edges.length <= 105
edges[i] = [ui, vi, si, musti]
0 <= ui, vi < n
ui != vi
1 <= si <= 105
musti is either 0 or 1.
0 <= k <= n
There are no duplicate edges.
Solutions
Solution 1: Binary Search + Union-Find
According to the problem description, the stability of a spanning tree is determined by the minimum strength edge in it. If a stability \(x\) is feasible, then for any \(y < x\), stability \(y\) is also feasible. Therefore, we can use binary search to find the maximum stability.
We first add all required edges into the Union-Find and record the minimum strength \(mn\) among them. If there is a cycle among the required edges, return \(-1\) directly. Then we add all edges into the Union-Find; if the final number of connected components is greater than \(1\), it means not all nodes can be connected, and we return \(-1\).
Next, we perform binary search in the range \([1, mn]\). We define a function \(\text{check}(lim)\) to check whether there exists a spanning tree with stability at least \(lim\). In the \(\text{check}\) function, we first add all edges with strength no less than \(lim\) into the Union-Find. Then we try to use the upgrade count to connect the remaining edges, with the condition that the edge strength is at least \(lim/2\) (since upgrading doubles the strength). If the final number of connected components in the Union-Find is \(1\), a spanning tree satisfying the condition exists.
The time complexity is \(O((m \times \alpha(n) + n) \times \log M)\), and the space complexity is \(O(n)\). Here, \(m\) is the number of edges, \(n\) is the number of nodes, and \(M\) is the maximum edge strength.