(2x - c)**2
has a minimum at x = c/2
. However, I have a much more complicated algebraic expression that I need to find the minimum of. And I need that minimum in terms of symbols, not an actual number. Specifically:
m(1-n)/(m(1-n) + (1-m)n) - x/(x(1-y) + (1-x)y)
where m = (x + (1-x)(1-c))
and n = (y + (1-y)(1-c))
for some constant c. x, y are both in [0, 1].
What would be a good way to go about finding this? I'd like to find value of x that maximizes this in terms of y and c. Even for y=0 would still be really great. I was trying to use Wolfram-Mathematica but it was not cooperating.
I got confused, your title says minimum and the text says maximum. You can indeed find a maximum :
..This takes a few minutes..
oddly, the values
{0,1/2,Infinity}
are correct, however thex,y
locations are not. It should bex==0 OR y==0 for c==1
andx==0 AND y==0 for c<1
.. EvidentlyMaximize
is not very good at describing a solution which is not unique.