Description cvc cvd

[neilwalton]

 
1 There seems to be a mismatch between the function and the help file:
 
Help file states:  capacity = demand + (cvd2 + cvc2) / waiting_time
Function calculates capacity = demand + ((cv_demand2 + cv_capacity2) / 2) * (factor / target_wait)
 
The * or / difference at the end is just rearranging,  but it’s the ‘/2’ on the sum of the squared coefficients of variation.  Should it have the /2 or not?
 
 
2 We should probably explain the coefficients of variation in the help files a bit better, for a future release.
 
Am I right to describe them as:  “the standard deviation of demand divided by the mean demand” and the same for capacity?  The default here is 1 for each, but I find that a bit hard to believe when I try and apply it to my data.  The more I think about it, is this because they arrivals are a Possion process, and the mean / variance both the rate {\lambda}?  That would mean, though that the sd would be \sqrt \lambda.  Am I a mile off here?
 
3 Applying this to my case:
 
I’m modelling a year’s data based on the previous year.  I know my annual demand and waiting list size, so I’m using the calc_relief_capacity function.
I’m not sure how the num_referrals argument works here and whether I should be using it, but it makes no difference when I do.  The cv_demand here defaults to 0 rather than 1.
Is this right?
 
1 There seems to be a mismatch between the function and the help file:
 
Help file states:  capacity = demand + (cvd2 + cvc2) / waiting_time
Function calculates capacity = demand + ((cv_demand2 + cv_capacity2) / 2) * (factor / target_wait)
 
The * or / difference at the end is just rearranging,  but it’s the ‘/2’ on the sum of the squared coefficients of variation.  Should it have the /2 or not?
 
 
2 We should probably explain the coefficients of variation in the help files a bit better, for a future release.
 
Am I right to describe them as:  “the standard deviation of demand divided by the mean demand” and the same for capacity?  The default here is 1 for each, but I find that a bit hard to believe when I try and apply it to my data.  The more I think about it, is this because they arrivals are a Possion process, and the mean / variance both the rate (?  That would mean, though that the sd would be .  Am I a mile off here?
 
3 Applying this to my case:
 
I’m modelling a year’s data based on the previous year.  I know my annual demand and waiting list size, so I’m using the calc_relief_capacity function.
I’m not sure how the num_referrals argument works here and whether I should be using it, but it makes no difference when I do.  The cv_demand here defaults to 0 rather than 1.
Is this right?