So in the first case, when talk about a plain circular ring, I assume you mean an annular ring, with a well defined inner radius and a different well defined out radius. With a positive charge at the center of the annular ring, positive charges will be repelled outward and negative charges attracted inward. Incidentally, not all the positive charge will go to the outside edge, since as more charge builds up there, it becomes energetically less favourable to move another positive charge there. In a 3-D conductor with 2-D surfaces, everything goes to one of the surfaces, but in a 2-D (or lower) conductor (with 1-D surfaces), there's not enough space on the surface, so you can have charge in the 'bulk' of the conductor, away from the edges, as long as the field everywhere in the conductor is zero. (References: [1] R. Friedberg, Am. J. Phys. 61, 1084 (1993), [2] D.J. Griffiths and Y. Li, Am. J. Phy. 64, 706, (1996))
With a mobius strip like the one in your illustration, Even if the concepts of inside and outside are not well defined, you can still talk about distance from the central point charge, and about electric field. Assuming that your mobius strip is made of a conductor, the charge will redistribute themselves until the electric field everywhere inside the material of the conductor is zero. For portions like the right hand portion of the strip in your picture, that means you will have a well defined separation of charge, because the local inner edge is definitely closer than the local outer edge. But on the left hand portion of the strip in your picture, both inner and outer edges are roughly the same distance from the center, so there will not be any separation of charge (there's no point to it, energetically). Being a 2-D conductor, not all the charge will have to be at the edge. So my rough impression is that you will have two 'circles/loops' of charge, one positive and one negative, which are well separated on the right, but converge and merge on the left.