Absolute Value

I am having an issue with finding the max value of positive and negative values.  I want it to always pick the highest integer but to have it's sign remain the same.  

For instance,

x=10, y=15, z=-20 are my numbers calculated that 

t:=550*[max(x,y,z)] -p is depended on.

 if I do r:= max(x,y,z) it will state that r=15 when I need it to pick the worse case for an equation it will be integrated in later on. 

if I do r:= max(IxI,IyI,IzI) it will state r=20 but not -20 cause of the absolute value.

Is there a different approach I can do where I wouldn't have to use solve for t for each single case and pick the max absolute value of all three?