s**9
发帖数: 207 | 1 How about a greedy algorithm? Output the greatest Pij, then delete row i and
column j; repeat for N times.
Proof that Pij is in one optimal solution: suppose there is an optimal
solution that doesn't include Pij. Then in the solution, bidder i is
assigned with an item k where Pik<=Pij. If Pik is replaced with Pij, the
solution is still optimal if Pij=Pik or is better than the original one if
Pij>Pik. Therefore, Pij must be included in a optimal solution. |
|
c***l
发帖数: 8 | 2 A DP algorithm based on the balance partition algorithm (can google find the
algorithm) for the first question.
大牛看看有什么问题.
#include
#include
#include
using namespace std;
extern int mindiff(int i, int A[]);
int main()
{
int A[5]={3, 5, 7, 11, 13};
int min;
min=mindiff(5,A);
cout<
}
int mindiff(int n, int A[])
{
int i, j;
int max=0;
int jmin;
int result;
double min;
double sum=0;
for(i=0;i<=n-1;i++) { sum+=A[... 阅读全帖 |
|
l****z
发帖数: 29846 | 3 By Stephen Schwartz
The Hartford Seminary has occupied a leading position among theological
faculties around the nation in accommodating Islam, particularly in its
radical forms, since appointing its first Muslim faculty member more than
twenty years ago. Today, that accommodation extends to the murderous regime
of Syrian dictator Bashar Assad.
Hartford maintains a highly questionable relationship with an Islamic
religious and Arabic-language school, the Al-Fatiha Islamic Academy in
Damascus, a... 阅读全帖 |
|
d*****a
发帖数: 70 | 4 哦，这样就make sense了
我觉得可以这么解，大体思路就是conditioning on first 3 steps：
Let 0 be tail and 1 be head, and
let pij = prob{X > Y| A gets i and B gets j in the first roll}, where i,j =
{0, or 1}.
Then, conditioning on first roll, for consistency, the order of conditioning
is always 00, 01, 10, 11
(1) p00 =
1/4 *p00
+ 1/4 *p01
+ 1/4 *p10
+ 1/4 *p11
Given first roll is 0 for A and 1 for B, conditioning on second roll:
(2) p01 =
1/4 *p00
+ 1/4 *[1/4 *p00... 阅读全帖 |
|
|
a**r
发帖数: 19 | 6 【 以下文字转载自 Statistics 讨论区 】
发信人: acar (心中的阳光), 信区: Statistics
标 题: Question about consistent estimator
发信站: BBS 未名空间站 (Sun Nov 6 06:09:21 2005), 转信
{Xn} is an ergodic Markov chain on Z+={0,1,2,...}
How to determine a consistent estimator of the transition
probabilities {Pij}?
Thanks! |
|
m*********g
发帖数: 646 | 7 另外第二题，我的答案 p(M>=N)=0.26
类似第一题方法，只不过A的结果和B的结果是独立事件。
下面求的是P（M>=N）的概率
note, pij 表示 P( desired |A有i个H in a row,B有j个H in a row)
列出来如下方程：
3/4p00=1/4p10+1/4p11+1/4p01
p10=1/4p00+1/4p01+1/2p_A
p11=1/2p2+1/4p02+1/4p00
p01=1/4p10+1/4p02+1/4p12+1/4p00
p02=1/2p_B+1/4p10+1/4p00
p12=1/2p_B+1/4p00+1/4p2
boundary 是p_A=0; p_B=1
6个方程有点多，解系统方程AX=B,得 0.26
还是等大牛们来指正吧。或者有更简便的方法。 |
|
