pull_group1_name = bilayer

Thanks for the answer. But putting bilayer as group 1 won't set it as the reference, right?

These are the parameters I am using for pulling:

pull = yes

pull_print_com1 = yes

pull_print_ref-value=yes

pull_print_components = yes

pull_ngroups =2

pull_ncoords =1

pull_group2_name = BILAYER

pull_group1_name = VESICLE

pull_coord1_groups = 1 2

pull_coord1_type = umbrella

pull_coord1_geometry = distance

pull_coord1_vec = 0.0 0.0 -1.0

pull_coord1_dim= N N Y

pull_coord1_rate = 0

pull_coord1_k = 100

pull_coord1_start = yes

I am not sure if I set group1= Bilayer and don't apply any restrains on the bilayer, then it will not move towards the vesicle. I just want to pull the vesicle in -Z direction (towards the bilayer). Is this code correct then? do you have any idea?

With the above code, you're pulling the bilayer towards the vesicle because you have the bilayer as group 2 and the vesicle as group 1, then defining the restraint groups as 1 2.  So the vesicle is the "reference" for the calculation. If you invert the definition/labeling of the groups, you'll be pulling the vesicle towards the bilayer.

Think more abstractly. The pull code re-write (sometime before 5.0, IIRC) means multiple biasing potentials can be applied simultaneously between any number of groups; you're no longer limited to just "reference" and "pulled group." What you should care about is the relative position of each group with respect to the other(s), e.g. the COM distance/reaction coordinate. If the bilayer moves but the vesicle still approaches it, you've accomplished what you wanted.

Don't get confused with the unique circumstances of the tutorial, which uses a relatively rigid group to (1) avoid distortion and (2) dramatically simplify the demonstration. There's no reason that a "reference" has to be a fixed point in space. For every action, there is an equal but opposite reaction...