Geometric Probability& Irwin–Hall Distribution几何概型

1. Core Knowledge of Geometric Probability

1.1 Definition & Essential Properties

Geometric probability is a fundamental continuous probability model, applicable to random experiments with infinitely many outcomes.

Two necessary and sufficient conditions (judgment criteria):

  • Infinity: The experiment has infinitely many possible outcomes.
  • Equal likelihood: Each outcome occurs uniformly and randomly with equal probability.

1.2 Geometric Probability vs Classical Probability

Probability ModelOutcome NumberEqual LikelihoodMeasure Type
Classical ProbabilityFiniteYesCounting number
Geometric ProbabilityInfiniteYesLength / Area / Volume / High-dimensional measure

1.3 Universal Probability Formula

$$P(A)=\frac{\text{Valid geometric measure of event }A}{\text{Total geometric measure of all possible outcomes}}$$

Classification of measures by dimension:

  • 1-dimensional (line/angle): length, angle
  • 2-dimensional (plane): area
  • 3-dimensional (space): volume
  • High-dimensional: n-dimensional volume (hypervolume)

2. Key Formula: Irwin–Hall Distribution

This is the core formula for solving the AMC ant problem, essentially the n-dimensional geometric probability volume formula.

2.1 Preconditions

Let \(X_1,X_2,…,X_n\) be independent and identically distributed (i.i.d.) random variables, uniformly distributed on the interval \((0,1)\), i.e., \(X_i \sim U(0,1)\).

Define the sum: \(S_n=X_1+X_2+…+X_n\)

Distribution function: \(F(x)=P(S_n \le x)\)

2.2 Piecewise Formula (Core Examination Range)

Only the first interval is required for this problem:

  • When \(0 \le x \le 1\): \(\boldsymbol{F(x)=\dfrac{x^n}{n!}}\)
  • When \(1 < x < n\): Complex segmented formula (not used in this problem)
  • When \(x \ge n\): \(F(x)=1\) (certain event)

2.3 Geometric Interpretation

The formula corresponds to the volume of ann-dimensional right simplex (n-dimensional tetrahedron) in the unit n-dimensional cube:

  • \(n=2, x=1\): Area of triangle \(=\dfrac{1^2}{2!}=\dfrac{1}{2}\)
  • \(n=3, x=1\): Volume of tetrahedron \(=\dfrac{1^3}{3!}=\dfrac{1}{6}\)

2.4 Critical Complementary Corollary

For \(0 \le x \le 1\):

$$P(S_n > x)=1-P(S_n \le x)=1-\frac{x^n}{n!}$$

3. Full Analysis of the AMC Ant Problem

3.1 Problem Rules Restatement

Amelia takes at most 3 steps. For each step \(n=1,2,3\):

  • Random time \(t_n \in (0,1)\), random distance \(x_n \in (0,1)\), all variables independent.
  • Stop rule: Stop during the nth step if total time exceeds 1 minute; stop after at most 3 steps.
  • Goal: Find the probability that the final position \( > 1\).

3.2 Three Mutually Exclusive Stopping Cases

All cases are disjoint and exhaustive: Case1 (1-step stop), Case2 (2-step stop), Case3 (3-step stop)

Case 1: Stop at Step 1 (\(t_1 > 1\))

Since \(t_1 \in (0,1)\), \(t_1 > 1\) is impossible. \(P_1=0\)

Case 2: Stop at Step 2 (\(t_1 \le 1, t_1+t_2 > 1\))

Time condition probability: \(P(t_1+t_2 > 1)=1-\dfrac{1^2}{2!}=\dfrac{1}{2}\)

Distance condition probability: \(P(x_1+x_2 > 1)=1-\dfrac{1^2}{2!}=\dfrac{1}{2}\)

Joint probability: \(P_2=\dfrac{1}{2} \times \dfrac{1}{2}=\dfrac{1}{4}\)

Case 3: Stop at Step 3 (\(t_1+t_2 \le 1\))

Time condition probability: \(P(t_1+t_2 \le 1)=\dfrac{1^2}{2!}=\dfrac{1}{2}\)

Distance condition probability: \(P(x_1+x_2+x_3 > 1)=1-\dfrac{1^3}{3!}=\dfrac{5}{6}\)

Joint probability: \(P_3=\dfrac{1}{2} \times \dfrac{5}{6}=\dfrac{5}{12}\)

3.3 Total Probability

$$P=P_1+P_2+P_3=0+\frac{1}{4}+\frac{5}{12}=\frac{2}{3}$$

4. Common Mistakes & Exam Tips

  • Formula Domain Restriction: \(\dfrac{x^n}{n!}\) is only valid for \(0 \le x \le 1\).
  • Impossible 1-step stop: \(t_n \in (0,1)\), so total time can never exceed 1 in the first step.
  • Complementary Priority: Calculate \(P(S_n>1)\) via complementary probability to avoid complex integral/geometry calculation.
  • Zero-measure boundary: Boundary lines/points have 0 measure, so \(\le\) and \( < \) have identical probability in geometric probability.

5. Core Summary

  1. The problem is a high-dimensional geometric probability problem, extended from basic 2D/3D geometric probability.
  2. Irwin–Hall formula unifies the volume calculation of n-dimensional uniform variable sum regions.
  3. Solve multi-step random process problems via classified discussion + mutually exclusive event probability addition.

Part 2: 中文授课讲义

1. 几何概型核心知识点

1.1 定义与两大核心性质

几何概型是连续型概率模型,适用于试验结果有无限多个的随机试验。

两大充要判定条件

  • 无限性:试验所有可能结果个数无限;
  • 等可能性:所有结果均匀随机出现、机会均等。

1.2 几何概型 vs 古典概型(核心区分)

概率模型结果个数等可能性计算测度
古典概型有限个满足个数计数
几何概型无限个满足长度/面积/体积/高维测度

1.3 通用概率公式

$$P(A)=\frac{\text{事件A对应的有效几何测度}}{\text{总试验区域对应的几何测度}}$$

维度对应测度

  • 一维(线段/角度):长度、角度
  • 二维(平面):面积
  • 三维(空间几何体):体积
  • 高维:n维超体积

2. 核心解题公式:Irwin–Hall 分布公式

该公式是本题解题关键,本质是n维几何概型体积通式,完美衔接初高中基础几何概型与高维概率。

2.1 适用前提

设 \(X_1,X_2,…,X_n\) 为相互独立、服从区间 \((0,1)\) 均匀分布的随机变量,记和 \(S_n=X_1+X_2+\dots+X_n\),分布函数 \(F(x)=P(S_n \le x)\)。

2.2 分段公式(考试必考区间)

仅需掌握本题适用的核心区间:

  • 当 \(0 \le x \le 1\) 时\(\boldsymbol{F(x)=\dfrac{x^n}{n!}}\)
  • 当 \(1 < x < n\):分段复杂公式(本题不涉及)
  • 当 \(x \ge n\):\(F(x)=1\)(必然事件)

2.3 几何意义(学生必懂)

公式对应n维单位立方体内,满足 \(X_1+X_2+\dots+X_n \le x\) 的n维直角单纯形(n维四面体)体积

  • 二维:2个变量和≤1,直角三角形面积 \(=\dfrac{1}{2!}=\dfrac{1}{2}\)
  • 三维:3个变量和≤1,正四面体体积 \(=\dfrac{1}{3!}=\dfrac{1}{6}\)

2.4 补集推论(解题神器)

在 \(0 \le x \le 1\) 范围内,求变量和大于定值的概率:

$$P(S_n > x)=1-\frac{x^n}{n!}$$

3. AMC蚂蚁难题完整解析

3.1 题意核心规则

蚂蚁最多爬行3步,每一步随机取耗时 \(t_n\in(0,1)\)、前进距离 \(x_n\in(0,1)\),所有变量相互独立;若第n步总耗时超过1分钟则停止,求停止时总位置大于1的概率。

3.2 三类互斥停止情况(全覆盖、无重叠)

情况1:第一步直接停止(\(t_1>1\))

因 \(t_1\in(0,1)\),不可能大于1,故 \(P_1=0\)。

情况2:第二步停止(\(t_1\le1,t_1+t_2>1\))

时间条件概率:\(P(t_1+t_2>1)=1-\dfrac{1^2}{2!}=\dfrac{1}{2}\)

距离条件概率:\(P(x_1+x_2>1)=\dfrac{1}{2}\)

联合概率:\(P_2=\dfrac{1}{2} \times \dfrac{1}{2}=\dfrac{1}{4}\)

情况3:走完三步停止(\(t_1+t_2\le1\))

时间条件概率:\(P(t_1+t_2\le1)=\dfrac{1^2}{2!}=\dfrac{1}{2}\)

距离条件概率:\(P(x_1+x_2+x_3>1)=1-\dfrac{1^3}{3!}=\dfrac{5}{6}\)

联合概率:\(P_3=\dfrac{1}{2} \times \dfrac{5}{6}=\dfrac{5}{12}\)

3.3 总概率计算

$$P=0+\frac{1}{4}+\frac{5}{12}=\frac{2}{3}$$

4. 高频易错点(课堂必讲避坑)

  • 公式定义域限制:\(\dfrac{x^n}{n!}\) 仅在 \(0\le x\le1\) 成立,超出区间失效;
  • 杜绝思维误区:第一步不可能超时,概率恒为0;
  • 优先用补集思想:高维区域求「和大于定值」,用1减去「和小于定值」的体积,大幅简化计算;
  • 几何概型边界无影响:边界线段、平面、单点测度为0,\(\le\) 和 \( < \) 概率完全相等。

5. 课堂终极总结

  1. 本题属于高维几何概型拔高题,是课内二维、三维几何概型的延伸;
  2. Irwin–Hall公式统一了所有低维、高维均匀变量和的区域测度计算;
  3. 多步骤随机概率问题,核心解法为分类讨论+互斥事件概率相加

Q1 2022 AMC 10B Problems/Problem 23

Ant Amelia starts on the number line at $0$ and crawls in the following manner. For $n=1,2,3,$ Amelia chooses a time duration $t_n$ and an increment $x_n$ independently and uniformly at random from the interval $(0,1).$ During the $n$th step of the process, Amelia moves $x_n$ units in the positive direction, using up $t_n$ minutes. If the total elapsed time has exceeded $1$ minute during the $n$th step, she stops at the end of that step; otherwise, she continues with the next step, taking at most $3$ steps in all. What is the probability that Amelia’s position when she stops will be greater than $1$?

$\textbf{(A) }\frac{1}{3} \qquad \textbf{(B) }\frac{1}{2} \qquad \textbf{(C) }\frac{2}{3} \qquad \textbf{(D) }\frac{3}{4} \qquad \textbf{(E) }\frac{5}{6}$

Q2 根据上面的题目修改成4步

Ant Amelia starts on the number line at $0$ and crawls in the following manner. For $n = 1, 2, 3, 4$, Amelia chooses a time duration $t_n$ and an increment $x_n$ independently and uniformly at random from the interval $(0,1)$. During the $n$-th step of the process, Amelia moves $x_n$ units in the positive direction, using up $t_n$ minutes. If the total elapsed time has exceeded $1$ minute during the $n$-th step, she stops at the end of that step; otherwise, she continues with the next step, taking at most $4$ steps in all. What is the probability that Amelia’s position when she stops will be greater than $1$?

中文:

蚂蚁 Amelia 从数轴上的原点 $0$ 出发。对于第 $n$ 步($n = 1, 2, 3, 4$),Amelia 独立且在区间 $(0,1)$ 内均匀随机地选择时间持续量 $t_n$ 和位置增量 $x_n$。在第 $n$ 步中,Amelia 向正方向移动 $x_n$ 个单位,消耗 $t_n$ 分钟。如果在第 $n$ 步期间总共消耗的时间超过了 $1$ 分钟,她就会在这一步结束时停止;否则继续下一步,总共最多走 $4$ 步。求 Amelia 停止时的位置大于 1 的概率。

Q3 修改范围(最多走 3 步,目标改为 1/2)

Ant Amelia starts on the number line at $0$ and crawls in the following manner. For $n = 1, 2, 3$, Amelia chooses a time duration $t_n$ and an increment $x_n$ independently and uniformly at random from the interval $(0,1)$. During the $n$-th step of the process, Amelia moves $x_n$ units in the positive direction, using up $t_n$ minutes. If the total elapsed time has exceeded $1$ minute during the $n$-th step, she stops at the end of that step; otherwise, she continues with the next step, taking at most $3$ steps in all. What is the probability that Amelia’s position when she stops will be greater than $1/2$ [OR: not greater than $1/2$]?

中文:

蚂蚁 Amelia 从数轴上的原点 $0$ 出发。对于第 $n$ 步($n = 1, 2, 3$),Amelia 独立且在区间 $(0,1)$ 内均匀随机地选择时间持续量 $t_n$ 和位置增量 $x_n$。在第 $n$ 步中,Amelia 向正方向移动 $x_n$ 个单位,消耗 $t_n$ 分钟。如果在第 $n$ 步期间总共消耗的时间超过了 $1$ 分钟,她就会在这一步结束时停止;否则继续下一步,总共最多走 $3$ 步。求 Amelia 停止时的位置大于 1/2(或不大于 1/2)的概率。

Q1 C2/3

Q2

蚂蚁可能在第 2、3、4 步结束时停下,利用 $1/n!$ 定理分三种情况:

  1. 在第 2 步停止(概率 $\frac{1}{4}$)
    • 时间超 1 分钟:$1 – \frac{1}{2!} = \frac{1}{2}$
    • 距离大于 1:$1 – \frac{1}{2!} = \frac{1}{2}$
    • 路径概率:$\frac{1}{2} \times \frac{1}{2} = \frac{1}{4}$
  2. 在第 3 步停止(概率 $\frac{5}{18}$)
    • 时间在第 3 步超 1 分钟:$\frac{1}{2!} – \frac{1}{3!} = \frac{1}{3}$
    • 距离大于 1:$1 – \frac{1}{3!} = \frac{5}{6}$
    • 路径概率:$\frac{1}{3} \times \frac{5}{6} = \frac{5}{18}$
  3. 在第 4 步停止(概率 $\frac{23}{144}$)
    • 前 3 步时间都没超 1 分钟:$\frac{1}{3!} = \frac{1}{6}$
    • 4 步总距离大于 1:$1 – \frac{1}{4!} = \frac{23}{24}$
    • 路径概率:$\frac{1}{6} \times \frac{23}{24} = \frac{23}{144}$

最终总合流 (Total Sum)

$$P = \frac{1}{4} + \frac{5}{18} + \frac{23}{144} = \frac{36 + 40 + 23}{144} = \frac{99}{144} = \frac{11}{16}$$

  • 最终答案 (Final Answer): $\frac{11}{16}$

Q3

如果问题是“大于 1/2” (Greater than 1/2)

阶梯式解答 (Solution Steps)

时间触发概率保持不变,利用 $\frac{h^n}{n!}$ 定理(此时 $h = 1/2$):

  1. 在第 2 步停止
    • 时间超 1 分钟:$\frac{1}{2}$
    • 距离大于 1/2:$1 – \frac{(1/2)^2}{2!} = 1 – \frac{1}{8} = \frac{7}{8}$
    • 路径概率:$\frac{1}{2} \times \frac{7}{8} = \frac{7}{16}$
  2. 在第 3 步停止
    • 时间在第 3 步才超(或到头停止):$\frac{1}{2}$
    • 距离大于 1/2:$1 – \frac{(1/2)^3}{3!} = 1 – \frac{1}{48} = \frac{47}{48}$
    • 路径概率:$\frac{1}{2} \times \frac{47}{48} = \frac{47}{96}$

最终总合流 (Total Sum)

$$P = \frac{7}{16} + \frac{47}{96} = \frac{42 + 47}{96} = \frac{89}{96}$$

  • “大于 1/2” 的最终答案: $\frac{89}{96}$

💡 方向 B:如果问题是“不大于 1/2” (Not greater than 1/2,即 $\le 1/2$)

阶梯式解答 (Solution Steps)

如果问的是“不大于 $1/2$”,其实就是方向 A 的对立事件(直接用 $1 – \frac{89}{96} = \frac{7}{96}$ 即可)。如果要写出独立的证明步骤,直接套用单纯形体积公式会极为漂亮:

  1. 在第 2 步停止
    • 时间超 1 分钟:$\frac{1}{2}$
    • 距离不大于 1/2:直接套用定理 $\frac{(1/2)^2}{2!} = \frac{1}{8}$
    • 路径概率:$\frac{1}{2} \times \frac{1}{8} = \frac{1}{16} = \frac{6}{96}$
  2. 在第 3 步停止
    • 时间没超或走到头:$\frac{1}{2}$
    • 距离不大于 1/2:直接套用定理 $\frac{(1/2)^3}{3!} = \frac{1}{48}$
    • 路径概率:$\frac{1}{2} \times \frac{1}{48} = \frac{1}{96}$

最终总合流 (Total Sum)

$$P = \frac{6}{96} + \frac{1}{96} = \frac{7}{96}$$

  • “不大于 1/2” 的最终答案: $\frac{7}{96}$

评论

Leave a Reply