what were you thinking when you uploaded this video?! everything sucks about it, have you yourself ever watched it??????
@muaazkhalid1517 жыл бұрын
Sound sucks...
@ryanjackson0x8 жыл бұрын
Captioning would be nice.
@jhl49348 жыл бұрын
MaxSAT random method : 0.5-approximation 6:20 (X1 or -X2) and (X2 or X3 or -X4) and (-X1 or -X3 or X4) 每個clause用 Yi代表 7:50 max Y1 + Y2 + ... + Ym sigma(Xi)(when clause's element is positive) + sigma(1-Xi)(when clause's element is negative) >= Yi , 1 <= j <= m Xi = {0,1} , 1 <= i <= n Yj = {0,1} , 1 <= j <= m ------------------- relax : Xi,Yi = [0,1] 15:15 Rounding stage set Xi to be true with probability Yi (Yi is LP 的最佳解) Xi will be set {0,1} Yi range [0,1] ------------------- proof 19:35 根據極限定理 . . .
@ccugraduatealgorithms260 Жыл бұрын
謝謝您的回覆呦
@jhl49348 жыл бұрын
Integer Programming is NP-Complete Linear Programming is P Integer Programming 3:50 目標函數 (objective function) : Maximize X1+X2 限制式 (constraints) : 4*X1 - X2 <=8 2*X1 + X2 <=10 X1 , X2 = 0 or 1 -------------------------- IP : 0 or 1 => {0,1} LP : 0 ~ 1 => [0,1] -------------------------- 21:25 weight vertex cover minimum total weight 每個邊就一個限制式 Minimize : 5*X1 + 1*X2 + 5*X3 + 10*X4 + 4*X5 + 6*X9 + 7*X2 subject to : X1 + X2 >=1 X1 + X3 >=1 . . . Xi = {0,1} relax : Xi = [0,1] if Xi <=0.5 ; Xi = 1 if Xi > 0.5 ; Xi = 0
@jhl49348 жыл бұрын
Randominzed Approximation Algorithm 3-SAT 在clause中隨機將原宿設成 F or T 那麼括號不滿足的機率是1/8 E[C] : 每一個括號被滿足的機率 * clause個數= 1-(1/8) * m = 7/8 * m 近似比 : C*/E[C] <= m/(7m/8) = 8/7 k-SAT 近似比 : 2^k / (2^k-1) Max-2-Cut 給一個graph,給一個cut可以砍掉最多邊。(將vertex分兩群) random : 每點隨機分群。 每個編成為cut機率為1/2 C*/E[C] = 2 Max-K-Cut 每個編成為cut機率為k-1/k C*/E[C] = k/(k-1) greedy : 每次一個點,如果鄰居大多都是A群,自己則設為B群。