20. (12 marks) (I) Let the daily production quantities of products $A$ and $B$ be $x$ and $y$ respectively, with corresponding profit $z$. Then we have
$$\left\{ \begin{array} { l }
2 x + 1.5 y \leq W \\
x + 1.5 y \leq 12 \\
2 x - y \geq 0 \\
x \geq 0 , \quad y \geq 0
\end{array} \right.$$
The objective function is $z = 1000x + 1200y$.
[Figure]Solution diagram 1 for Question 20 [Figure]Solution diagram 2 for Question 20 [Figure]Solution diagram 3 for Question 20When $W = 12$, the planar region represented by (1) is shown in Figure 1, with three vertices $A ( 0,0 ) , B ( 2.4,4.8 ) , C ( 6,0 )$ respectively. Rewriting $z = 1000 x + 1200 y$ as $y = - \frac { 5 } { 6 } x + \frac { z } { 1200 }$, when $x = 2.4 , y = 4.8$, the line $l : y = - \frac { 5 } { 6 } x + \frac { z } { 1200 }$ has maximum $y$-intercept, maximum profit $Z = z _ { \max } = 2.4 \times 1000 + 4.8 \times 1200 = 8160$. When $W = 15$, the planar region represented by (1) is shown in Figure 2, with three vertices $A ( 0,0 ) , B ( 3,6 ) , C ( 7.5,0 )$ respectively. Rewriting $z = 1000 x + 1200 y$ as $y = - \frac { 5 } { 6 } x + \frac { z } { 1200 }$, when $x = 3 , y = 6$, the line $l : y = - \frac { 5 } { 6 } x + \frac { z } { 1200 }$ has maximum $y$-intercept, maximum profit $Z = z _ { \max } = 3 \times 1000 + 6 \times 1200 = 10200$. When $W = 18$, the planar region represented by (1) is shown in Figure 3, with four vertices $A ( 0,0 ) , B ( 3,6 ) , C ( 6,4 ) , D ( 9,0 )$ respectively. Rewriting $z = 1000 x + 1200 y$ as $y = - \frac { 5 } { 6 } x + \frac { z } { 1200 }$, when $x = 6 , y = 4$, the line $l : y = - \frac { 5 } { 6 } x + \frac { z } { 1200 }$ has maximum $y$-intercept, maximum profit $Z = z _ { \max } = 6 \times 1000 + 4 \times 1200 = 10800$. Thus the distribution of maximum profit $Z$ is
| $Z$ | 8160 | 10200 | 10800 |
| $P$ | 0.3 | 0.5 | 0.2 |
Therefore, $E ( Z ) = 8160 \times 0.3 + 10200 \times 0.5 + 10800 \times 0.2 = 9708$. (II) From (I), the probability that daily maximum profit exceeds 10000 yuan is $p _ { 1 } = P ( Z > 10000 ) = 0.5 + 0.2 = 0.7$. By the binomial distribution, the probability that at least one day out of 3 days has maximum profit exceeding 10000 yuan is $p = 1 - \left( 1 - p _ { 1 } \right) ^ { 3 } = 1 - 0.3 ^ { 3 } = 0.973$.