Facility Location
Consider the problem of locating a warehouse to serve a number of service locations. The design variable is the location of the warehouse, , while the service locations are given by the vector , .
Minimizing the maximum distance
|
One possible objective function for this problem involves the maximum distance from the warehouse to any location. This is the problem
which can be cast as the SOCP
|
This can be coded up with the CVX applet:
CVX syntax
% pts is a 2 x n matrix of n points
cvx_begin
variable x(2,1);
variable t(n,1);
minimize( max(t) )
subject to
for i = 1:n,
t(i) >= norm(x-pts(:,i));
end
cvx_end
Minimizing the average distance
|
Often, the average distance is a good measure of the transportation costs involved. This leads to the problem
which can be cast as the SOCP
|
This can be coded up with the CVX applet:
CVX syntax
% pts is a 2 x n matrix of n points
cvx_begin
variable x(2,1);
variable t(n,1);
minimize( max(t) )
subject to
for i = 1:n,
t(i) >= norm(x-pts(:,i));
end
cvx_end
|