Poser 9 /Pro2012 have some improvement for dynamic hair, so I thought I'd give it a new try. Just rendering free worked fine. Shaders need some tuning, but OK.

One of the advantages of dynamic hair is that in theory it can react to hats and head-bands. I assume there can be issues with the definition of inside and outside of a hat geometry, so I used a 'moniker', a pipe (from the mphing props that come with Poser) with large wall thickness, the lower rim and inside wall of the pipe resembling the inside of the hat, the outer diameter was large so there can be no ambiguity about the position of a vertex beig inside or outside the pipe volume. The pipe volume was clar of the hair at start of animation. During the animation the pipe inside was slowly closed on to the head, but the end contition was chosen so that there was ample space between the skull and the moniker for the hair to find a place.

For the experiment I took a curly hair (Adorana's Baronie hair) with 40 vertices per hair.  

I found the hair reacted to the moniker to some extent: the hair volume at the end of the animation had a clear cut-in where the edge of the moniker was, but still there was substantial intersection between hair and moniker, way more than could occur from under-sampling (the hair passing trough the cylinder only for the distance between two vertices that themselves are outside the 'hat'.)  

Is there anything I can do about this: Use a finer mesh for the moniker maybe, or use a moniker with a rounded edge? Any ideas anyone? 

Sounds like you will need to use magnets to finish up the process of hiding the hair within the hat.

Since your pseudo-hat has some of the required effect, I wonder if you could make a morph based on that and dial it to >1?

This doesn't work on dynamic cloth - morphs only work pre-simulation - so I suspect not, not without a lot of fiddling anyway.  

Bob may be on to something!

There is a Dynamics Parameter for each hair group simulation. I quite often dial it to 0.6 (default=1.0) or so to give the effect of some hairspray. I wonder if you can dial it greater than 1? That would also be useful for making a rather timid simulation somewhat wilder without having to redo the sim.

First try lowering your bend resistance. Possibly the root stiffness also since this is a hat. You might try increasing the vertexes but use a small group for that because this will up your sim time greatly.

You can also try using the morph brush. I don't remember if it will work on draped hair. If it doesn't, you can use Cage's bake deformation script to back the draping before using the tool. You might also want the sticky addon that prevents the morph tool from switching what is selected. It works better that the shift button when using the morph tool. I don't remember who wrote it though.

i maybe talking crazy, but what if you took the hat into the hair room and just grew some hair off it, then run the sim on the hair only? you dont need a full head. just an idea

@acadia: I am afraid so, but it would do away with the advantage and be as cumbersome as transmapped hair. I have not given up the hope yet!

@EnglishBob: That may be an idea. Spawn Morph Target will not work I am afraid, but I do not know whether dynamic hair would be exported in a .obj an could be read as a morph target.

@CaptainMARC: That is another good idea; thank you; will try this and revert.

@RedPhantom: Cage's bake Deformation script. I do not know that one, but if it does what the name suggests it may be useful. Will sort that out.

@MysticEagle: That is plan B indeed. It would work for closed and non-transparant hats only, and I would need to work out the hair style!. Thank you anyway. 

Thank you all for your responses. you gave me some good ideas. Will revert on this. (Need to do something these days ;-) )

 

Here is baron hair with a hat. What I did was hide all the hair groups that that stuck out through the hat more than from the bottom. Some wouldn't have shown at all. I enlarged the hat just enough that the hair wasn't poking though and ran a simulation using the settings shown. The only groups I used were hinten 1, Hinten2 rechts1 and links1. Oh I also moved the 2 side hair groups along the x axis because they were poking though the face some where the bends in the kink were.

Here's progress report! 

See the above. I used all the hair groups from the baron hair. A lot is hidden inside the 'turban' I made her (Antonia) wear with the hair. For practical applications of course I could make some invisible, move and use magnets, but the name of the game here is how to constrain hair. Other solutions are postfix recourse. 

The thing on het head is the 'mould' I used this time: A transformed torus. In fact it is the outer torus, because I used a second torus inside this one to make the 'pressure' on the hair more intense. During the simulations I noted the hair, when squeezed, at some moment jumps trough the surface of the mould and then, once passed, there appears to be no holding back anymore. Apparently, once the process has taken the 'penalty' asociated with  a hair vertex entering illegal space, there is no further penalty if it drifts away there further. By adding a second mould inside the first one, I added a new penalty. I can not give a weight to the penalty function in the collision. I assume this is something internal hard coded in Poser. I can not find it in the .cr2 and PoserPython does not give any handles either. Logic has it however that vilolating more constraints gives more penalty. Will try agian later and see what happens if more consecutive constraints are imposed, either as back-up to the final surface, or as 'waymaker', preceeding the final surface. 

A second conclusion I made is that it is better to actually let the hat come down on the head than to 'shrink' the hat onto the head. The crux seems to be that the process lacks methods to find a place for the hair when 'caught' between head and hat. The available input parameter make me assume the structural model of the hair contains only bending and maybe torsion. Strain, and certainly 'buckling' appear to be not represented in the model. Hence the best way to move the hair aside would be a 'sliding' load, much like 'combing' your hair with your hand.            

The 'dynamics' parameter for the hair groups indeed can be set to values higher than 1.0.

Capturing the morph I have not investigated yet.   

To be continued!