思路：单调递增栈 + k 控制删除次数。高位越小整体越小，遇更小数字时弹出栈顶大数（仅当 k0）；栈空且当前为 0 则跳过（避免前导零）；若遍历完 k 仍0，从末尾再删 k 位。
В России ответили на имитирующие высадку на Украине учения НАТО18:04
，更多细节参见旺商聊官方下载
Медведев вышел в финал турнира в Дубае17:59
The N-closest or N-best dithering algorithm is a straightforward solution to the N-candidate problem. As the name suggests, the set of candidates is given by the closest palette colours to the input pixel. To determine their weights, we simply take the inverse of the distance to the input pixel. This is essentially the inverse distance weighting (IDW) method for multivariate interpolation, also known as Shepard’s method. The following pseudocode sketches out a possible implementation: