Proc OPTMODEL for pool selection

[hckadali]

Hi,

     I am trying to select 200 points from a pool of 1000 points which are nearest to the line say y = x using least squares. and my constraint is that the average of the values of the 200 points should be below a given value. How can i formulate this model using Proc optmodel?

I am not sure how i write the objective function in this case. I thought it should be like

min f = {sum i in 1..200} (x - y)**2

but in the above objective function, there is nothing to minimize, it just calculates the sum of least squares of first 200 obs. I am confused on how to tackle this. Plz help.

Thank you!

[PGStats]

"the average of the values of the 200 points should be below a given value" ? The average of what value, x, y, SSQ?

[hckadali]

There is one more column, say 'z'.

Average of  (y/z) of those 200 points should be less than some given value.

If i am correct in representing in SAS OPT model notation:

con constraint: {sum i in 1..200} (y/z) / 200  <= 0.60