Find the probability that all sets are satisfied. I am facing this problem while writing a code.
Assume that we have the following sets:
$$S_1= \{P_1, P_2, P_3, P_9\}$$
$$S_2= \{P_2, P_4, P_6, P_9\}$$
$$S_3= \{P_6, P_7, P_8, P_9\}$$
$$S_4= \{P_3, P_5, P_7, P_9\}$$
Let $T$ be a set that contains some $P_i$ where $1 \le i \le 9$ such that the probability of $P_i$ being in $T$ is $K_i$ (in my case, $K_i \in \{0, 0.2\}$ but let's just say $K_i$). What is the probability $A$ that $T$ contains at least an element from each $S_i$?
I couldn't find a simple formula to solve this. I am not sure if it is easier to approach this if we try to find $1 - A$ first.
 A: As a first approach, one can examine all $2^9$ possible outcomes. Each outcome is like $o_t =(0,0,1,0,1,0,0,1,0)$  where $t=1,\dots,2^9$. The probability of $o_t$ is always $P(o_t)=\prod_i K_i^{o_{t,i}}(1-K_i)^{1-o_{t,i}}$. For each $o_t$ you can automatically check whether it satisfies the condition, i.e. $o_t\in A$. Since all elementary events are independent, you can sum it as
$$P(A)=\sum_{o_t\in A} P(o_t)$$.
I know this is not an elegant closed form, but I think that it is implementable easily.
A: Let $L_i=1-K_i$.
Let $M_i$ be the event that $S_i$ is entirely missing.
Then, using Inclusion-Exclusion, 
$$A = 1...\\-Pr(M_1)-Pr(M_2)-Pr(M_3)-Pr(M_4)\\+Pr(M_1\cap M_2)+Pr(M_1\cap M_3)+...\\
-Pr(M_1\cap M_2\cap M_3)-Pr(M_1\cap M_2\cap M_4)-...\\
+Pr(M_1\cap M_2\cap M_3\cap M_4)\\
=1-L_1L_2L_3L_9-L_2L_4L_6L_9-L_6L_7L_8L_9-L_3L_5L_7L_9\\
+L_1L_2L_3L_4L_6L_9+L_1L_2L_3L_6L_7L_8L_9+...\\
-L_1L_2L_3L_4L_6L_7L_8L_9-...\\
+L_1L_2L_3L_4L_5L_6L_7L_8L_9
$$
There will be four terms with one $M_i$, six terms with two $M_i$, four terms with three $M_i$ and one with all four $M_i$.  That makes 16 terms including the initial $1$. 
