📚 Combinations of Random Variables | 随机变量的组合
In Edexcel A-Level Mathematics, the topic of combinations of random variables is essential for understanding how to model the sum, difference, and linear scaling of independent quantities. This concept allows you to determine the expectation and variance of new random variables formed from existing ones, with widespread applications in probability and statistics. Mastery of these techniques is crucial for tackling problems involving the normal, Poisson, and binomial distributions, as well as sampling theory.
在Edexcel A-Level数学中，随机变量的组合这一主题对于理解如何对独立量的和、差以及线性缩放进行建模至关重要。这一概念可以帮助你确定由现有随机变量形成的新随机变量的期望和方差，并在概率与统计中有着广泛应用。熟练掌握这些技巧对于解决涉及正态分布、泊松分布、二项分布以及抽样理论的问题至关重要。
1. Introduction to Combinations of Random Variables | 随机变量组合简介
A linear combination of random variables X and Y takes the form aX + bY + c, where a, b, and c are constants. In Edexcel Statistics, you will often be given the means, variances, and the key condition of independence, and then asked to compute the expectation and variance of the combination. This topic forms the foundation for many areas, including hypothesis testing and confidence intervals.
随机变量 X 和 Y 的线性组合具有 aX + bY + c 的形式，其中 a、b 和 c 为常数。在 Edexcel 统计学中，你通常会得到均值、方差以及独立这一关键条件，然后被要求计算该组合的期望和方差。这一主题为假设检验、置信区间等许多领域奠定了基础。
It is important to distinguish between the expectation operator, which is always linear, and the variance, which requires careful handling of covariance. Under independence, the covariance term vanishes, making calculations straightforward.
区分期望算子与方差至关重要：期望总是线性的，而方差则需要小心处理协方差。当随机变量独立时，协方差项为零，计算变得直接。
2. Expectation of a Linear Combination | 线性组合的期望
(Y) + c. This linearity of expectation is a powerful tool, allowing you to break down complex expressions into manageable parts.
对于任何随机变量 X 和 Y，无论它们是否独立，线性组合的期望都遵循一条简单的规则：E(aX + bY + c) = aE(X) + bE(Y) + c。期望的线性性质是一个强大的工具，它使你能将复杂的表达式分解为易于处理的部分。
E(aX + bY) = a E(X) + b E(Y)
he constant term simply adds to the total, and scalar multiplication applies directly to the individual means.
例如，若 E(X) = 10 且 E(Y) = 6，则 2X - 3Y + 5 的期望为 2(10) - 3(6) + 5 = 20 - 18 + 5 = 7。请注意，常数项直接加到总和上，而标量相乘则直接作用于单个均值。
3. Variance of a Linear Combination for Independent Variables | 独立变量线性组合的方差
Var(Y). The constant c has no effect on variance, and it is vital to note that the coefficients are squared.
当 X 和 Y 独立时，线性组合的方差以一种特定的方式合并单个变量的方差：Var(aX + bY + c) = a²Var(X) + b²Var(Y)。常数 c 对方差没有影响，并且务必注意各系数被平方。
Var(aX ± bY) = a²Var(X) + b²Var(Y) (independent)
) = 9, then Var(3X - 2Y) = 3²(4) + (-2)²(9) = 9(4) + 4(9) = 36 + 36 = 72. The sign in front of Y does not matter because the coefficient is squared.
一个常见的考试错误是忘记当变量独立时，方差总是相加的，即使对于差值也是如此。例如，若 Var(X) = 4 且 Var(Y) = 9，则 Var(3X - 2Y) = 3²(4) + (-2)²(9) = 9(4) + 4(9) = 36 + 36 = 72。Y 前面的符号无关紧要，因为系数被平方了。
4. Sums of Independent Random Variables | 独立随机变量之和
E(Y) and Var(S) = Var(X) + Var(Y). This is extremely useful when dealing with total outcomes, such as the total weight of several items or the total number of events occurring in independent processes.
当 a = 1 且 b = 1 时，线性组合简化为和 S = X + Y。对于独立变量，期望和方差简化为：E(S) = E(X) + E(Y) 且 Var(S) = Var(X) + Var(Y)。这在处理总结果时极为有用，例如若干物品的总重量或独立过程中发生事件的总数。
If you have more than two independent random variables, say X₁, X₂, …, Xn, the sum T = X₁ + X₂ + … This forms the basis for the distribution of sample totals and the Central Limit Theorem.
如果你有两个以上的独立随机变量，比如 X₁、X₂、…、Xn，其和 T = X₁ + X₂ + … + Xn 的期望为 E(T) = ΣE(Xi)，方差为 Var(T) = ΣVar(Xi)。这为样本总和的分布和中心极限定理奠定了基础。
5. Differences of Independent Random Variables | 独立随机变量之差
X) + Var(Y). It is a typical pitfall to subtract variances, but the variance of the difference is the sum of the variances when X and Y are independent.
差值 D = X - Y 是一个 a = 1、b = -1 的线性组合。利用上述公式，E(D) = E(X) - E(Y)，Var(D) = 1²Var(X) + (-1)²Var(Y) = Var(X) + Var(Y)。一个典型的陷阱是以为方差可以相减，但实际上当 X 和 Y 独立时，差值的方差是方差之和。
ariability of both components contributes positively to the variability of the difference.
这一性质可能看起来有违直觉，但其原因在于减去 Y 等价于加上 -Y，而 Var(-Y) = Var(Y)。因此，两个分量的变异性都对差值的变异性产生正向贡献。
6. Linear Combinations of Normal Variables | 正态变量的线性组合
If X and Y are independent normal random variables, any linear combination aX + bY + c also follows a normal distribution. Specifically, if X ~ N(μ₁, σ₁²) and Y ~ N(μ₂, σ₂²), then aX + bY + c ~ N(aμ₁ + bμ₂ + c, a²σ₁² + b²σ₂²). This reproductive property makes the normal distribution particularly amenable to statistical analysis.
如果 X 和 Y 是独立的正态随机变量，那么任何线性组合 aX + bY + c 也服从正态分布。具体而言，若 X ~ N(μ₁, σ₁²) 且 Y ~ N(μ₂, σ₂²)，则 aX + bY + c ~ N(aμ₁ + bμ₂ + c, a²σ₁² + b²σ₂²)。这种再生特性使得正态分布特别适合进行统计分析。
For example, suppose the length of a bolt and the length of a nut are independent normal variables with means 50 mm and 20 mm, and variances 0.25 mm² and 0.09 mm², respectively. The total length of the bolt plus nut is L ~ N(50+20, 0.25+0.09), i.e., N(70, 0.34). You can then use standard normal tables to find probabilities such as P(L > 70.5).
例如，假设螺栓长度和螺母长度是独立的正态变量，均值分别为 50 mm 和 20 mm，方差分别为 0.25 mm² 和 0.09 mm²。螺栓加螺母的总长度 L ~ N(50+20, 0.25+0.09)，即 N(70, 0.34)。然后你可以用标准正态分布表来求概率，如 P(L > 70.5)。
7. Sums of Poisson Variables | 泊松变量的和
Independent Poisson random variables are also reproducible under addition. If X ~ Po(λ₁) and Y ~ Po(λ₂) are independent, then X + Y ~ Po(λ₁ + λ₂). This property is very useful for modelling the total number of events from two independent Poisson processes, such as calls arriving at two separate call centres.
独立的泊松随机变量在加法下也具有再生性。若 X ~ Po(λ₁) 且 Y ~ Po(λ₂) 独立，则 X + Y ~ Po(λ₁ + λ₂)。这一性质对于对来自两个独立泊松过程的事件总数进行建模非常有用，例如两个不同呼叫中心的来电数量。
However, be careful: the difference X - Y is not a Poisson variable, and linear combinations like 2X + 3Y do not generally follow a Poisson distribution. Only the simple sum of independent Poissons retains the Poisson form.
但要小心：差值 X - Y 不是泊松变量，像 2X + 3Y 这样的线性组合通常也不服从泊松分布。只有独立泊松变量的简单求和才能保持泊松分布的形式。
8. Combinations Involving Binomial Variables | 涉及二项变量的组合
For binomial distributions, the sum of independent variables is again binomial only when the probability of success p is the same. If X ~ B(n₁, p) and Y ~ B(n₂, p) are independent, then X + Y ~ B(n₁ + n₂, p). This result is often used when pooling data from different trials with the same underlying success rate.
对于二项分布，只有当成功概率 p 相同时，独立变量之和才仍然是二项分布。如果 X ~ B(n₁, p) 且 Y ~ B(n₂, p) 独立，那么 X + Y ~ B(n₁ + n₂, p)。当合并具有相同基础成功率的不同试验数据时，这一结果经常被用到。
can still be found using the linear combination rules: E(X+Y) = n₁p₁ + n₂p₂, Var(X+Y) = n₁p₁q₁ + n₂p₂q₂.
如果成功概率不同，总数就不服从简单的二项分布。在这种情况下，仍然可以用线性组合的规则来求总和的期望和方差：E(X+Y) = n₁p₁ + n₂p₂，Var(X+Y) = n₁p₁q₁ + n₂p₂q₂。
9. Application of the Central Limit Theorem | 中心极限定理的应用
m (CLT) allows us to approximate the distribution as normal, even if the original variables are not normal. For a sum S = ΣXi, where each Xi has mean μ and variance σ², S is approximately N(nμ, nσ²) for large n.
当处理大量独立同分布 (i.i.d.) 随机变量的和或均值时，中心极限定理 (CLT) 允许我们将其分布近似为正态分布，即使原始变量并非正态。对于总和 S = ΣXi，其中每个 Xi 的均值为 μ、方差为 σ²，当 n 很大时，S 近似服从 N(nμ, nσ²)。
150 g and standard deviation 20 g is approximately N(100×150, 100×20²) = N(15000, 40000), and we can estimate probabilities like P(S > 15200).
这在质量控制和民意调查中特别方便。例如，从一个均值为 150 g、标准差为 20 g 的总体中随机抽取 100 个苹果的总重量，近似服从 N(100×150, 100×20²) = N(15000, 40000)，我们可以估计像 P(S > 15200) 这样的概率。
10. Worked Examples and Common Mistakes | 示例解析与常见错误
Example 1: Let X ~ N(10, 4) and Y ~ N(5, 1) be independent. Find P(2X - Y > 18).
We have E(2X - Y) = 2(10) - 5 = 15, Var(2X - Y) = 4(4) + 1 = 17. So 2X - Y ~ N(15, 17). The z-score is (18 - 15)/√17 ≈ 0.7276, giving a probability of 1 - 0.766 = 0.234 (from tables).
示例1：设 X ~ N(10, 4) 与 Y ~ N(5, 1) 独立。求 P(2X - Y > 18)。
我们有 E(2X - Y) = 2(10) - 5 = 15，Var(2X - Y) = 4(4) + 1 = 17。因此 2X - Y ~ N(15, 17)。z 分数为 (18 - 15)/√17 ≈ 0.7276，查表得概率为 1 - 0.766 = 0.234。
Common Mistake 1: ar(X - Y) = Var(X) - Var(Y). Always remember that variance adds for independent variables.
Common Mistake 2: Forgetting to square the coefficients in variance calculations.
Common Mistake 3: Using the addition rule for Poisson variables when the variables are not independent or when a linear combination other than a simple sum is used.
常见错误1：假设 Var(X - Y) = Var(X) - Var(Y)。务必记住，对于独立变量，方差是相加的。
常见错误2：在方差计算中忘记将系数平方。
常见错误3：在变量不独立或使用非简单求和的线性组合时，套用泊松变量的加法规则。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图