这是有图片的.md

MachineLearningMathBasic/2.8 Moore-Penrose伪逆-李涛.md

@@ -1,8 +1,7 @@
 # 2.8 Moore-Penrose伪逆-李涛
 
-![-w797](media/15669588348534/15669601763364.jpg)
-![-w803](media/15669588348534/15669602014898.jpg)
-
+![](https://tva1.sinaimg.cn/large/006y8mN6ly1g6fjbz5i93j318a0qiag2.jpg)
+![](https://tva1.sinaimg.cn/large/006y8mN6ly1g6fjdlbpduj318m0iun0r.jpg)
 ```python
 import scipy as SP
 

MachineLearningMathBasic/6.3 凸集与凸集分离定律-李涛.md

@@ -0,0 +1,45 @@
+# 6.3 凸集与凸集分离定律-李涛
+![](https://tva1.sinaimg.cn/large/006y8mN6ly1g6fjg1lhx6j30r20rk421.jpg)
+![](https://tva1.sinaimg.cn/large/006y8mN6ly1g6fjg1f5udj30pe13sdje.jpg)
+![](https://tva1.sinaimg.cn/large/006y8mN6ly1g6fjg19xnxj30p604o0u7.jpg)
+```
+from cvxpy import *
+import cvxpy as cvx
+
+```
+```
+problem :
+minimize (x-y)^2
+subject to
+        x+y=1
+        x-y>=1
+```
+```
+from cvxpy import *
+# Create two scalar optimization variables.
+# 在CVXPY中变量有标量(只有数值大小)，向量，矩阵。
+# 在CVXPY中有常量(见下文的Parameter)
+x = Variable() #定义变量x,定义变量y。两个都是标量
+y = Variable()
+# Create two constraints.
+#定义两个约束式
+constraints = [x + y == 1,
+              x - y >= 1]
+#优化的目标函数
+obj = Minimize(square(x - y))
+#把目标函数与约束传进Problem函数中
+prob = Problem(obj, constraints)
+prob.solve()  # Returns the optimal value.
+print("status:", prob.status)
+print("optimal value", prob.value) #最优值
+print("optimal var", x.value, y.value) #x与y的解
+# Replace the objective. 不同的目标函数，相同的约束
+prob2 = cvx.Problem(cvx.Maximize(x + y), prob.constraints)
+print("optimal p1 value", prob2.solve())
+
+# Replace the constraint (x + y == 1).#不同的约束，相同的目标函数
+constraints = [x + y <= 3] + prob.constraints[1:] #注意：此处是列表相加，prob.constraints[1:]取prob的约束集中
+#从第二个开始的所有约束。
+prob2 = cvx.Problem(prob.objective, constraints)
+print("optimal P2 value", prob2.solve())
+```