91Ó°ÊÓ

In quality control, the binomial distribution is often used to model the number of defective items in a sample. The main idea is that each item in the sample can only be in one of two states: defective or not defective.
For a scenario like the one given in the exercise, where a random sample of 500 transistors is taken from a shipment, the binomial distribution can help us compute the likelihood of finding a certain number of defective transistors.
Here, the binomial distribution is defined by two parameters: the sample size ( = 500) and the probability of a single item being defective (p = 0.04).
The formula to calculate the probability of finding exactly k defective transistors in our sample is:
\[ P(D = k) = \binom{n}{k} p^k (1-p)^{n-k} \] where:

• \(\binom{n}{k}\) is the binomial coefficient representing the number of ways to choose k defective items from n samples.
• p^k is the probability of k defective items occurring.
• (1-p)^{n-k} is the probability of the remaining items being non-defective.

In our specific case, we need to compute the probability for all values of k from 0 to 9 (as shipments are rejected if they contain 10 or more defective transistors).

Quality control is vital in manufacturing settings to ensure that products meet certain standards and specifications. Acceptance sampling is a technique used in quality control to decide whether to accept or reject a batch of products based on a sample. The goal is to identify whether the batch is of acceptable quality without inspecting every item.
In our problem, a quality control engineer takes a random sample of 500 transistors out of a shipment of 50,000. The engineer will reject the entire shipment if 10 or more transistors in the sample are defective.
The idea is to balance the risk of accepting bad quality goods (producers' risk) against the cost of inspecting every item or the risk of unnecessarily rejecting good quality goods (consumers' risk). The probability calculations using binomial and normal distribution help us assess this balance. If the probability of accepting a shipment with a high defect rate is very low, the acceptance policy might be deemed too strict.
This can lead to higher costs due to the unnecessary rejection of batches that meet acceptable quality standards. Conversely, if the threshold is set too high, poor quality goods may be accepted, leading to downstream problems. Finding a suitable policy goes a long way in maintaining both production efficiency and product quality.

In many real-world applications, especially when dealing with large sample sizes, computing probabilities directly from the binomial distribution can be complex. That's where the normal approximation to the binomial distribution comes in handy.
The normal approximation makes it easier to compute probabilities by approximating the binomial distribution with a normal distribution. This method is particularly useful when both np and n(1-p) are greater than 5.
In our problem:

• np = 500 * 0.04 = 20
• n(1-p) = 500 * 0.96 = 480

Since both these values are comfortably greater than 5, the normal approximation is applicable. The mean (\text{mean} = np) and standard deviation (\text{std} = \text{sqrt}(np(1-p))) of the binomial distribution can be used to define the normal distribution: \[ D \thicksim N(20, \text{sqrt}(19.2)) \] The continuity correction factor (0.5) is used because a continuous distribution is being used to approximate a discrete distribution. So, calculating: \[ P(D < 9.5) = P\bigg(Z < \frac{9.5 - 20}{4.38}\bigg) \] converts our problem to finding a standard normal value, which helps determine the probability.The result here indicates the shipment acceptance probability is very low (about 0.82%). Therefore, the policy primarily results in rejecting shipments even if the actual defect rate isn't too far from acceptable, which might need reassessment to avoid excessive rejections.