题目
Problem
Kris works in the mailroom of a large company and is responsible for all the letters sent by the company. The weights of letters sent by the company, W grams, have a normal distribution with mean 165 g and standard deviation 35 g.
(a) Estimate the proportion of letters sent by the company that weigh less than 120 g.
(3)
Kris splits the letters to be sent into 3 categories: heavy, medium and light, with 31 of the letters in each category.
(b) Find the weight limits that determine medium letters.
(4)
A heavy letter is chosen at random.
(c) Find the probability that this letter weighs less than 200 g.
(3)
Kris chooses a random sample of 3 letters from those in the mailroom one day.
(d) Find the probability that there is one letter in each of the 3 categories.
(3)
解答
(a)
解法一
思路
展开
先标准化,再查 Φ(z)。因为 120 在平均数左边,所以最后要取左尾概率。
答题过程
展开
P(W<120)=P(Z<35120−165)=P(Z<−1.2857…).
Using symmetry,
P(Z<−1.2857…)=1−P(Z<1.2857…).
From tables,
P(Z<1.29)≈0.9015.
Therefore
P(W<120)≈1−0.9015=0.0985.
(b)
解法一
思路
展开
三类各占三分之一,所以 medium 是中间三分之一。边界对应标准正态的下三分之一点和上三分之一点,即 z≈±0.43。
答题过程
展开
The medium letters are between the lower one-third point and the upper one-third point.
Using the percentage points of the normal distribution,
z≈0.43.
So the two limits are found from
35x−165=±0.43.
Hence
x=165±0.43(35).
Therefore
x=165±15.05.
The weight limits are approximately
150 g and 180 g.
(c)
解法一
思路
展开
已知选到的是 heavy letter,所以分母是 P(W>180)=31。要小于 200,则分子是 180<W<200。
答题过程
展开
Using the upper limit for medium letters from part (b), a heavy letter has W>180.
Therefore
P(W<200∣W>180)=P(W>180)P(180<W<200).
Now
35200−165=1.
So
P(W<200)=P(Z<1)=0.8413.
Also P(W≤180)≈32, so
P(180<W<200)=0.8413−32.
Hence
P(W<200∣W>180)==310.8413−320.524….
Therefore the probability is approximately
0.524.
解法二
思路
展开
条件概率补集法。
已知选到的是 heavy letter(即 W>180)。我们要求 P(W<200∣W>180),可以先求其对立事件(补集法则):
P(W⩾200∣W>180)=P(W>180)P(W⩾200∩W>180)
由于 200>180,事件 W⩾200 完全包含在 W>180 之中。因此,交集部分直接简化为单个变量概率:
P(W⩾200∩W>180)=P(W⩾200)
由此可得:
P(W⩾200∣W>180)=P(W>180)P(W⩾200)=1/31−P(W<200)=3×P(W⩾200)
通过求补集,分子部分直接化简为单尾概率,省去了复杂的区间概率差 P(180<W<200) 的计算,在代数上极具优雅性,且避免了分式中循环小数 2/3 的截断误差。
答题过程
展开
Use the complement rule for conditional probability:
P(W<200∣W>180)=1−P(W⩾200∣W>180).
Since 200>180, the event W⩾200 is a subset of W>180. Therefore:
P(W⩾200∩W>180)=P(W⩾200).
By definition of conditional probability:
P(W⩾200∣W>180)=P(W>180)P(W⩾200).
Standardise 200:
P(W⩾200)====P(Z⩾35200−165)P(Z⩾1)1−0.84130.1587.
Since P(W>180)=31, we have:
P(W⩾200∣W>180)===1/30.15873×0.15870.4761.
Thus:
P(W<200∣W>180)==1−0.47610.5239.
Therefore the probability is approximately
0.524.
(d)
解法一
思路
展开
每一类的概率都是 31。三封信要恰好一封属于每类,先写某一个顺序的概率,再乘以 3! 个排列。
答题过程
展开
For one particular order, for example heavy, medium, light, the probability is
31⋅31⋅31=271.
There are
3!=6
possible orders.
Therefore
P(one in each category)=6⋅271=92.