I understand that exponential distribution can be modeled as a function $$ f(x) = \lambda e^{-\lambda x} u(x) $$

Suppose a scenario where I want to observe the lifespan of an object and $T$ is the time in year belonging to an exponential distribution, with an expected value of $4$ years. May I know how would I go about calculating the $\lambda$ for $t = 3, 4, 5$?

Thank you.

  • $\begingroup$ And $u(x)$ is...? Is it the Heaviside step function? $\endgroup$ Mar 23, 2018 at 15:39
  • $\begingroup$ @DavidG.Stork: I suspect $u=1_{[0,\infty)}$. $\endgroup$
    – copper.hat
    Mar 23, 2018 at 15:41
  • 3
    $\begingroup$ If $T$ is exponentially distributed, then $ET = {1 \over \lambda} = t$, so set $\lambda = {1 \over t}$. Is this what you are looking for? Look at en.wikipedia.org/wiki/Exponential_distribution $\endgroup$
    – copper.hat
    Mar 23, 2018 at 15:43

1 Answer 1


Suppose a particular kind of electronic components have lifetimes $X_1 \sim \mathsf{Exp}(\text{rate}=\lambda = 1/4).$ Then the average time to failure is $\mu = E(X_1) = 1/\lambda = 4.$ That is, the average component lasts for about four years. The density function is $f_1(t) = .25e^{-.25t},$ for $t > 0.$ You can show by calculus that $\mu = \int_0^\infty tf_1(t)\,dt = 1/4.$

The probability such a component will fail within half a year is $F_1(.5) = 1 = e^{-.25(1/2)} = 0.1175,$ where $F_1(t) = 1 - e^{-\lambda t}.$ In R statistical software the computation can be done in two ways, where pexp is the name of the CDF.

 1 - exp(-.25*.5)
 ## 0.1175031

 pexp(.5, .25)
 ## 0.1175031

In R statistical software, you can generate lifetimes of 1000 components, and make a histogram of them as shown below. The R function rexp generates a random sample and dexp is an exponential density function.

x1 = rexp(1000, .25)
hist(x1, prob=T, ylim=c(0,.25), col="skyblue2", main="Hist of 1000 obs from EXP(.25)")
curve(dexp(x, .25), add=T, lwd=2, col="blue")
abline(v = .5, lwd=2, col="red")

enter image description here

Now, if you want to model lifetimes over a four year time period, you could use $$X_4 \sim \mathsf{Exp}(\lambda = 4(.25) = 1).$$ Then $E(X_4) = 1.$ If there are .25 failures on average in one year then there is 1 on average in a four-year period.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.