density function and expected value of binomial-like distribution

196 Views Asked by At

Suppose a computer needs 20 memory chips to work, and there is a 0.1 probability that a given chip is defective.

Part 1: If $X$ denotes the number of chips to be ordered to ensure that 20 chips work, what is the probability density function of the random variable $X$?

Part 2: What is the expected value of that density function?

My attempt: In my opinion, the problem wasn't worded as clearly as it could have been. I interpreted Part 1 to ask for the density function $f(x)$, where $f(x)$ is the probability that exactly 20 of the $x$ chips work. Under this interpretation, I previously thought the density function should be $f(x)={x \choose 20}0.9^{20}0.1^{x-20}$. However, this density function does not appear to sum to one. I've tried numerous other potential density functions, and I can't seem to find one that sums to one. I would greatly appreciate any advice on this.

1

There are 1 best solutions below

1
On BEST ANSWER

Your answer to the first part is not correct, because if we conceptualize the ordering process as comprising successive orders of one chip at a time until the $20^{\rm th}$ non-defective chip is obtained, then we stop precisely when that $20^{\rm th}$ good chip is obtained, hence the last ordered chip is always good. Consequently, there are only $\binom{x-1}{19}$ ways of arranging the previous $19$ good chips among the $x-1$ previous chips ordered before the last good one is obtained, hence $$\Pr[X = x] = \binom{x-1}{19} (0.9)^{20} (0.1)^{x-20}, \quad x = 20, 21, 22, \ldots.$$ This is a negative binomial distribution with parameters $r = 20$ and $p = 0.9$.

To find $\operatorname{E}[X]$, it suffices to observe by construction that $$\sum_{x=r}^\infty \Pr[X = x] = \sum_{x=r}^\infty \binom{x-1}{r-1} p^r (1-p)^{x-r} = 1.$$ Next, consider $$x \binom{x-1}{r-1} = \frac{x!}{(r-1)! \, (x-r)!} = r \cdot \frac{x!}{r! \, (x-r)!} = r \binom{x}{r}.$$ Thus $$\begin{align*} \operatorname{E}[X] &= \sum_{x=20}^\infty x \Pr[X = x] \\ &= \sum_{x=20}^\infty x \binom{x-1}{19} (0.9)^{20} (0.1)^{x-20} \\ &= \frac{20}{0.9} \sum_{x=20}^\infty \binom{(x+1)-1}{21 - 1} (0.9)^{21} (0.1)^{(x+1)-21} \\ &= \frac{200}{9} \sum_{x=21}^\infty \binom{x-1}{21-1} (0.9)^{21} (0.1)^{x-21} \\ &= \frac{200}{9}, \end{align*}$$ where in the last step we recognize the sum as being that of a negative binomial random variable with parameters $r = 21$, $p = 0.9$ over its support, so it has value $1$. In general, we see that for parameters $r$ and $p$, $$\operatorname{E}[X] = \frac{r}{p}$$ for this particular parametrization of the negative binomial distribution.