Sliding window minimum algorithm

This is a homework problem. Let A[] is an array of integers and integer K -- window size. Generate array M of minimums seen in a window as it slides over A. I found an article with a solution for this problem but did not understand why i开发者_如何学Pythont has O(n) complexity. Can anybody explain it to me?

---

This tends to catch people out. You would think it would take O(N^2) time since you reason adding takes O(N) time and you have O(N) elements. However, realise each element can only be added once and removed once. So in total it takes O(N) to slide over the whole array A.

This yields an amortised efficiency of O(1) every time you move the sliding window on by one element. In other words, the average time it takes to move the sliding window by one element is O(1).