[AMPL 2630] Re: obtaining minimum between two variables

On Jul 6, 2:24 pm, plozmar <lozano_pal...@...> wrote:
I assume X and Y (and hence Z) are nonnegative.  If you know upper
bounds for X and Y, and can cope with binary variables (making the
model a MILP), then yes.

var B {NODES} binary;
param MX {NODES};  # MX[n] is a valid a priori upper bound on X[n]
param MY {NODES};  # MY[n] is a valid a priori upper bound on Y[n]

...

s.t. Zdef1 {n in NODES}: Z[n] >= X[n] - MX[n]*B[n];
s.t. Zdef2 {n in NODES}: Z[n] >= Y[n] - MY[n]*(1 - B[n]);

If Z[n] is not used anywhere in the model other than the objective,
this is sufficient; the solver will set B[n] based on whichever value
produces the smaller value of Z[n].  If Z[n] is used in other
constraints, so that there is a danger the solver might "want" to pick
a value of Z[n] greater than the min of X[n] and Y[n], then you also
need the following:

s.t. Xsmaller {n in NODES}: X[n] <= Y[n] + MX[n]*B[n];        # X <= Y
if B = 0
s.t. Ysmaller {n in NODES}: Y[n] <= X[n] + MY[n]*(1 - B[n]);  # Y <= X
if B = 1
s.t. Zmax1 {n in NODES}: Z[n] <= X[n] + MX[n]*B[n];
s.t. Zmax2 {n in NODES}: Z[n] <= Y[n] + MY[n]*(1 - B[n]);

/Paul
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "AMPL Modeling Language" group.
To post to this group, send email to ampl@...
To unsubscribe from this group, send email to ampl+unsubscribe@...
For more options, visit this group at http://groups.google.com/group/ampl?hl=en
-~----------~----~----~----~------~----~------~--~---