Poisson Distribution Modelling: Typical Application Scenarios | 泊松分布建模的典型应用场景

📚 Poisson Distribution Modelling: Typical Application Scenarios | 泊松分布建模的典型应用场景

The Poisson distribution is a cornerstone of discrete probability in Edexcel A-Level Further Mathematics. It describes the number of rare, independent events occurring in a fixed interval of time or space, and it appears in fields as diverse as physics, telecommunications, insurance and biology. In this article we explore the conditions under which a Poisson model is valid and examine the most important application scenarios you need to know for the examination.

泊松分布是爱德思 A-Level 进阶数学中离散概率的基石。它描述在固定时间或空间区间内稀有独立事件的发生次数，广泛应用于物理、通信、保险和生物学等领域。本文将探讨泊松模型成立的条件，并系统梳理你需要在考试中掌握的最典型应用场景。

---

1. Definition and Key Facts | 定义与核心性质

The Poisson distribution with parameter λ (lambda) is a discrete distribution defined on the non-negative integers. The probability of observing exactly r events is given by the probability mass function below, where e ≈ 2.71828 is Euler’s number, λ > 0 is the average rate of occurrence, and r! denotes the factorial of r.

参数为 λ（lambda）的泊松分布是定义在非负整数上的离散分布。恰好观察到 r 次事件的概率由以下概率质量函数给出，其中 e ≈ 2.71828 是自然对数的底数，λ > 0 是事件的平均发生率，r! 表示 r 的阶乘。

P(X = r) = e⁻λ × λʳ / r! , r = 0, 1, 2, …

The distribution is written X ~ Po(λ). Its most striking property is that the mean and the variance are equal; a single parameter λ controls the entire shape of the distribution.

该分布记为 X ~ Po(λ)。它最引人注目的性质是均值与方差相等；单一参数 λ 就完全决定了分布的形状。

E(X) = Var(X) = λ

---

2. The Conditions for a Poisson Model | 泊松模型的适用条件

A Poisson model is not simply assigned to any count data; it is only valid when four conditions are satisfied. First, events must occur independently, meaning that one event happening does not change the probability of another. Second, events occur at a constant average rate within the interval considered. Third, the probability of two or more events occurring in a very small interval is negligible, so events actually occur singly. Fourth, we are counting events over a fixed interval of time, length, area or volume.

泊松模型不是随手套用在任何计数数据上的；它只有在满足四个条件时才成立。第一，事件之间必须独立，即一次事件的发生不会改变另一次事件发生的概率。第二，在所考察的区间内，事件以恒定的平均速率发生。第三，在非常小的区间内发生两次或更多事件的概率可以忽略不计，因此事件实际上是逐个发生的。第四，我们是在固定的时间、长度、面积或体积区间内对事件进行计数。

• Independence: events must not influence each other. 独立性：事件之间互不影响。
• Constant rate: λ is fixed for the interval. 恒定速率：区间内 λ 保持固定。
• Rare events: two events cannot occur simultaneously. 稀有性：两个事件不能同时发生。
• Fixed interval: we count over a fixed window. 固定区间：我们在固定窗口内计数。

---

3. Physics — Radioactive Decay | 物理——放射性衰变

One of the classic exam scenarios is radioactive decay. In a sample of unstable isotopes, each nucleus decays independently of the others, and for a source whose half-life is long compared with the counting time, the average number of decays per second is effectively constant. A Geiger-Müller tube placed near the source therefore records a count of ionising particles that follows a Poisson distribution.

经典考点场景之一是放射性衰变。在不稳定同位素样本中，每个原子核都独立于其他原子核发生衰变。只要半衰期远长于计数时间，每秒的平均衰变数就近似恒定。因此，放置在放射源附近的盖革-米勒计数管记录到的电离粒子数服从泊松分布。

For example, if the average count rate is 12 particles per minute, the probability of recording exactly 15 particles in one minute is found by substituting λ = 12 and r = 15 into the formula.

例如，若平均计数率为每分钟 12 个粒子，则在一分钟内恰好记录到 15 个粒子的概率可通过将 λ = 12、r = 15 代入公式求得。

P(X = 15) = e⁻¹² × 12¹⁵ / 15! ≈ 0.0724

---

4. Call Centres and Queueing | 呼叫中心与排队论

Call centres are the most frequently used context for Poisson modelling in examinations. Incoming calls arrive at a switchboard at a certain average rate, for instance 5 calls per minute during the busy hour. Individual customers dial independently, and in a short time interval two simultaneous calls are very unlikely. The number of calls arriving in one minute is therefore modelled as X ~ Po(5).

呼叫中心是考试中最常见的泊松建模背景。来电以某一平均速率到达交换机，例如在繁忙时段平均每分钟 5 通电话。各位顾客是独立拨号的，且在很短的时间区间内，两通电话同时到达的可能性非常小。因此，一分钟内到达的电话数可建模为 X ~ Po(5)。

Typical questions ask you to find the probability of exactly 4 calls, or at most 2 calls. Using λ = 5:

典型问题会要求你求恰好接到 4 通电话的概率，或至多 2 通电话的概率。取 λ = 5：

P(X = 4) = e⁻⁵ × 5⁴ / 4! ≈ 0.1755

P(X ≤ 2) = e⁻⁵(1 + 5 + 25/2) = 18.5e⁻⁵ ≈ 0.1247

---

5. Road Traffic and Transport | 道路交通与运输

The number of vehicles passing a fixed point on a quiet road in a given time interval is another standard example. Cars arrive independently as long as traffic is free-moving and there is no congestion or platooning caused by traffic lights. If the average flow is 30 vehicles per hour, then the count in a 20-minute interval is Po(10), while the count in a 10-minute interval is Po(5).

在给定时间区间内通过某一固定路段的车辆数也是一个标准例子。只要交通处于自由流状态，没有由红绿灯引起的拥堵或车队成串现象，车辆到达就是相互独立的。若平均流量为每小时 30 辆，则 20 分钟区间内的车辆数为 Po(10)，而 10 分钟区间内的车辆数为 Po(5)。

This scaling of λ with the length of the interval is an important property: for a Poisson process, λ is proportional to the size of the interval. The same idea applies to gaps between vehicles, which follow the exponential distribution, a close partner of the Poisson process.

这种 λ 随区间长度按比例缩放的性质非常重要：对于泊松过程，λ 与区间大小成正比。同样的思想也适用于车辆之间的车头时距——它服从指数分布，而指数分布是泊松过程的亲密伙伴。

---

6. Insurance and Risk Management | 保险与风险管理

Insurance companies use the Poisson distribution to model the number of claims arising from a large portfolio of policies. For example, a life insurer with a very large number of policyholders might model the number of death claims in a month as Poisson, because deaths are rare, independent events with a roughly constant rate over a short period. Similarly, the number of car accident claims received in a day by a regional office is often assumed to be Poisson.

保险公司使用泊松分布为

Published by TutorHao | A-Level 进阶数学 Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）