Certain automated equipment needs part replacement on a random basis. This can be described by a Poisson distribution that has a mean of one part every other hour. When a breakdown occurs, a technician diagnoses the problem, and then goes to the parts counter to obtain the appropriate part. The average time for a clerk to obtain the part is 30 minutes because of the need to check various parts catalogs and retrieve the part from storage. Assume service time can be approximated by exponential distribution. Clerk time represents a cost of $30 per hour per clerk. Equipment down time and technician time represent $500 per hour. What number of clerks would be optimal for minimizing the total cost?