Because we are optimizing an average and the window size is 'at least k', this becomes a binary search on answer problem with prefix sum transformation. It is not simple sliding window due to varying window size and negative numbers.

Although a min heap works, the optimal approach is binary search on answer. The matrix's sorted structure gives a monotonic count of elements <= mid.

This is actually an Eulerian path problem in a directed graph. The lexical order requirement requires a priority queue. It is NOT standard topological sort.

This is a classic 0/1 knapsack (subset sum) DP problem. Greedy fails because local choices do not guarantee global partition.

Because movement cost varies (weights), BFS is invalid. This becomes a shortest path problem with weighted edges → Dijkstra.

Negative numbers break sliding window monotonicity. The correct approach uses prefix sums and a monotonic deque to maintain increasing prefix indices.

This is an interval DP problem. You must compute palindrome substrings and build DP on partition positions.

This is a weighted graph problem where we maximize probability instead of minimize distance. Use Dijkstra with a max heap.

Because window size is fixed (exactly k), use fixed sliding window. Kadane is for variable-length maximum subarray.

In a directed graph, cycle detection is done via topological sort or DFS with recursion stack. Union-Find works mainly for undirected graphs.