Introduction to the Exponential Distribution
Hello! Welcome to your study notes on the Exponential Distribution. This topic is a key part of your Optional application 2 – statistics module. While it might look intimidating with its \(e\) terms and integrals, it’s actually one of the most logical and "real-world" distributions you'll encounter.
Simply put, if the Poisson distribution counts how many events happen in a certain time, the Exponential distribution measures the time between those events. If you've ever waited for a bus or a text message, you've experienced the exponential distribution in action!
1. When to Use the Exponential Distribution
Before we look at the formulas, we need to know when this model is appropriate. The exponential distribution is used to model the time or distance between independent events occurring at a constant average rate (\(\lambda\)).
Key Conditions (SF1):
To use this distribution, the events must be:
- Independent: One event happening doesn't change the probability of the next one happening.
- Random: Events occur singly and cannot happen at exactly the same time.
- Constant Rate: The average number of events per unit time, denoted by \(\lambda\) (lambda), stays the same.
The Poisson Connection (SF4):
There is a beautiful link between these two distributions: If the number of events follows a Poisson distribution, the time between those events follows an Exponential distribution.
Example: If a shop receives an average of \(\lambda = 10\) customers per hour (Poisson), then the time you wait between customers arriving follows an Exponential distribution with the same \(\lambda\).
Quick Review: Remember that while the Poisson variable \(X\) is discrete (you can't have 2.5 customers), the Exponential variable \(X\) is continuous (you can wait 2.5 minutes).
2. The Probability Density Function (PDF)
The Probability Density Function, \(f(x)\), tells us the "height" of the curve at any point \(x\). For the exponential distribution, the curve starts high and decays towards zero.
The Formula (SF1):
For a random variable \(X \sim \text{Exp}(\lambda)\):
\(f(x) = \lambda e^{-\lambda x}\) for \(x \ge 0\)
\(f(x) = 0\) for \(x < 0\)
What does the graph look like?
Imagine a slide that starts at \(\lambda\) on the y-axis and curves downwards towards the x-axis, never quite touching it. This shows that short waiting times are much more likely than very long waiting times.
3. The Cumulative Distribution Function (CDF)
The Cumulative Distribution Function, \(F(x)\), is much more useful for solving exam problems. It tells you the probability that the waiting time is less than or equal to a certain value \(x\).
The Formula (SF1):
\(F(x) = P(X \le x) = 1 - e^{-\lambda x}\)
Calculating "Greater Than" Probabilities (SF2):
Often, questions ask for the probability that you wait longer than a certain time. This is even simpler:
\(P(X > x) = 1 - F(x) = e^{-\lambda x}\)
Memory Aid:
Think "L" for Less than (\(1 - e\)) and "M" for More than (Just the \(e\)).
\(P(X > x) = e^{-\lambda x}\) (The "tail" of the distribution).
Example Problem:
The time between calls to a helpdesk follows an exponential distribution with an average of 4 calls per hour. Find the probability that the next call arrives within 15 minutes.
1. Find \(\lambda\): 4 calls per hour.
2. Convert time to match \(\lambda\): 15 minutes = 0.25 hours.
3. Use \(F(x)\): \(P(X \le 0.25) = 1 - e^{-4 \times 0.25} = 1 - e^{-1} \approx 0.632\).
4. Mean, Variance, and Standard Deviation
You need to know these three properties of the exponential distribution (SF3). They are surprisingly simple!
The Mean (Expected Value):
\(E(X) = \mu = \frac{1}{\lambda}\)
Analogy: If you get 10 buses per hour, you expect to wait \(\frac{1}{10}\) of an hour (6 minutes) between them.
The Variance:
\(\text{Var}(X) = \sigma^2 = \frac{1}{\lambda^2}\)
The Standard Deviation:
\(\text{SD}(X) = \sigma = \sqrt{\frac{1}{\lambda^2}} = \frac{1}{\lambda}\)
Did you know? In an exponential distribution, the Mean and the Standard Deviation are exactly the same! If a question tells you the mean is 5, you automatically know the standard deviation is 5.
5. Required Proofs (SF3)
For AQA Further Maths, you need to be able to prove the mean and variance. Don't worry if these seem tricky; they just use Integration by Parts.
Proof for the Mean, \(E(X)\):
Recall that \(E(X) = \int_{0}^{\infty} x f(x) dx\).
\(E(X) = \int_{0}^{\infty} x (\lambda e^{-\lambda x}) dx\)
Using Integration by Parts where \(u = x\) and \(\frac{dv}{dx} = \lambda e^{-\lambda x}\):
1. \(\frac{du}{dx} = 1\) and \(v = -e^{-\lambda x}\)
2. \(\int u v' = [uv] - \int v u'\)
3. \(E(X) = [-xe^{-\lambda x}]_{0}^{\infty} - \int_{0}^{\infty} -e^{-\lambda x} dx\)
4. As \(x \to \infty\), \(-xe^{-\lambda x} \to 0\). At \(x=0\), it is \(0\).
5. \(E(X) = 0 + [\frac{-1}{\lambda} e^{-\lambda x}]_{0}^{\infty}\)
6. \(E(X) = 0 - (0 - \frac{1}{\lambda}) = \frac{1}{\lambda}\).
Key Takeaway: To prove the mean, you integrate \(x f(x)\). To prove the variance, you first find \(E(X^2)\) by integrating \(x^2 f(x)\) and then use \(\text{Var}(X) = E(X^2) - [E(X)]^2\).
6. Common Mistakes to Avoid
- Mixing up Units: Always make sure \(\lambda\) and your time \(x\) are in the same units (e.g., both in hours or both in minutes).
- Wrong \(\lambda\): Remember that \(\lambda\) is the rate (e.g., 5 per hour), while the mean is \(1/\lambda\) (0.2 hours). If the question says "the mean time is 10," then \(\lambda = 1/10\).
- Forgetting \(1 - \dots\): For \(P(X < x)\), it is \(1 - e^{-\lambda x}\). For \(P(X > x)\), it is just \(e^{-\lambda x}\). Students often flip these!
Summary Table
Parameter: \(\lambda\) (Rate)
PDF \(f(x)\): \(\lambda e^{-\lambda x}\)
CDF \(F(x)\): \(1 - e^{-\lambda x}\)
Mean: \(1/\lambda\)
Variance: \(1/\lambda^2\)
Context: Time/distance between random, independent events.