Amortized analysis

• 聚合分析
• 核算法
• 势能法
  ${\hat{c}}_i-c_i=credit=\Phi (D_i)-\Phi (D_{i-1})$
  $\sum {\hat{c}}_i=\sum c_i+\Phi (D-i)-\Phi (D_{i-1})$

合并两个skew heap摊还时间为O(logN)

exercises

C
A light child has at most half weight of its parents. Thus if there are k light nodes not including x along the path from x to y, $w(y)\le w(x)/2^k$ ( $k\le log\frac{w(x)}{w(y)}$ )
$\Phi$ = the total number of right heavy nodes it contains
By lemma 2, any path in a skew heap contains only O(logn) light nodes
Any heavy node on such a path is converted from right to left by causing a drop of one in the potential
Two heaps $h_1$ and $h_2$, containing $n_1$ and $n_2$ items
$n=n_1+n_2$
By lemma 1, the total number of light nodes is at most $2\lfloor logn\rfloor -1$
Let $k_1$ and $k_2$ be the number of heavy nodes on the right path of $h_1$ and $h_2$, and $k_3$ be the number of nodes that become right heavy children of nodes on the merge path.
By lemma 2, $k_3\le \lfloor logn\rfloor$
the number of node on the merge path is at most
$c_i\le 2+\lfloor log{n}_1\rfloor +k_1+\lfloor log{n}_2\rfloor +k_2\le 1+2\lfloor logn\rfloor +k_1+k_2$
The increase in the potential caused by the merge is $\Delta \Phi =k_3-k_2-k_1=\lfloor logn\rfloor -k_1-k_2$
This the amortized is at most $3\lfloor logn\rfloor +1$

https://blog.csdn.net/Woolseyyy/article/details/51517446