It is indeed possible. I'm getting close, but I need help the rest of the way.
I'm able to find the constants for one of the constraints using reduced cost and shadow price from the sensitivity report.
Lets redefine our problem so it becomes clearer:
Max Z = 5x1 + 4x2 + 3x3 + 4x4
a11*x1 + a12*x2 + a13*x3 + a14*x4 <= 40
a21*x1 + a22*x2 + a23*x3 + a24*x4 <= 30
Let y1 and y2 be the shadow price for constraints 1 and 2.
Let r1..r4 be the reduced cost for x1..x4.
Let c1..c4 be the constants for the max problem (here 5, 4, 3, 4).
Then:
r1 = c1 - (a11*y1 + a21*y2)
r2 = c2 - (a12*y1 + a22*y2)
and so on.
(Source page 80)
Since the shadow price is 0 for constraint 1 (y1=0) this gives
a21 = 6, a22=4, a23=4, a24=4
The second constraint then becomes
6*x1 + 4*x2 + 4*x3 + 4*x4 <= 30
How can I find the constants for the first constraint?
(According to the exam solution, it's supposed to be a11..a14 = 4, 6, 4, 2)
This is an exam problem, and there is no other given context. Not that it matters anyway; all excel sensitivity analysis are in the same format.
– tsorn Aug 01 '15 at 16:25