Probability of detecting at least two defects

462 Views Asked by At

I am struggling with the correctness of my solution. This is the problem: Product produced in a factory has a probability of 0,27 to be defective. Factory control detects a defect in a probability of 0,99 (it means it won't detect a defective product in a probability of 0,01). Question: what is the probability, that control between 5 products will find not less than 2 defective ones?

My solution: As I understood the question, they ask the probability of at least two items will be defective.

It means:

P(at least two) = 1 - P( no defects found) - P(one defect found)

I calculated P(no defects found) that way:

f = 0.27 (probability of producing defective part)

t = 1 - f = 0.73 (probability of producing a good part)

b1 = 0.99 (probability of detecting a defected product)

b2= 1 - b1 = 0.01 (probability of detecting a defected product)

Probabilities of P(no defects found) start there:

A0 = $t^5$ (all 5 parts are not defective)

A1 = $t^4$ * k * b2 (one defective, non found)

A2 = $t^3$ * $k^2$ * $b2^2$

A3 = $t^2$ * $k^3$ * $b2^3$ (three defective, non found)

...

A5 = $k^5$ * $b2^5$

P(one defect found) starts there:

B0 (no need to duplicate)

B1 = $t^4$ * k * b2 (one defective part, only one defect found)

B2 = $t^3$ * $k^2$ * b2 * b1 (two defective parts, only one defect found)

B3 = $t^2$ * $k^3$ * $b2^2$ * b1 (three defective parts, only one defect found)

...

B5 = $k^5$ * $b2^4$ * b1 (five defective parts, only one defect found)

So, the P(zero) is equal to A0 + A1 + ... A5 = 0.2080767583...

and P(one found) is equal to B1 + B2 + ...+ B5 = 0.07619029792

And P(at least two)= 1 - P(zero) - P(one found) = 0.7157329438

I don't know, is my solution correct and answer too, maybe there is another faster way to calculate these possibilites

1

There are 1 best solutions below

7
On BEST ANSWER

I found your analysis very difficult to wade through. So, my (somewhat inferior) response will simply show how I would attack the problem.

I am making some assumptions about how the problem constraints are to be interpreted. If I have misinterpreted the problem, please advise by leaving a comment, following this answer.


I am assuming that :

  • The probability that a specific item is defective is $(0.27).$

  • Assuming that a specific item is defective, the probability of the defect being detected is then $(0.99)$.

  • There is no possibility of a false positive. That is, if an item is not defective, then the testing will never falsely report that the item is defective.

  • Therefore, the probability that any item is reported as defective is $(0.27) \times (0.99).$

  • Testing of a specific item for a defect is independent of testing a separate item for a defect. Therefore (for example), the probability of detecting a defect on two consecutive items is $~\displaystyle \left[ ~(0.27) \times (0.99) ~\right]^2.$


For simplicity of notation, let $~p = (0.27) \times (0.99),~$ and let $~q = (1-p).$

So, $~p~$ equals the probability that a specific item will be reported as defective, and $~q~$ equals the complementary probability that the item will not be reported as defective.

Assuming that exactly $~5~$ items are tested,
for $~k \in \{0,1,2,3,4,5\},~$
let $~f(k)~$ denote the probability that exactly $k$ items are detected as defective.

Then, the desired computation is

$$1 - f(0) - f(1).$$

Then,

$$f(0) = \binom{5}{0}p^0q^{5-0} = q^5$$

and

$$f(1) = \binom{5}{1}p^1q^{5-1} = 5pq^4.$$


$\underline{\text{Final Computation}}$

$p = (0.27) \times (0.99).$
$q = 1-p.$

The probability of at least two defects being detected, out of five items tested, is

$$1 - q^5 - 5pq^4.$$