1. No. The the camber in a member has absolutely nothing to do with its bending capacity. However, any adverse affects due to heating the member to achieve the camber should be carefully observed.
I thought one of the benefits of precambering, if done correctly, was to induce a residual compressive stress in the lower flange/member so that as the part deflected it transfered the compressive stress to a tensile one, thus increasing the capacity.
1. Precambering does not reduce deflection. The total deflection will remain the same. Only the relative deflection will change. The purlin will start out at some precambered distance above the reference centerline, and will deflect to some distance below the reference line. The point of deflection below the reference line will be less than the non-cambered purlin. However, the total deflection stays the same. T
he point to cambering the purlin would be to camber it an amount equal to the expected sustained dead load deflection which will occur. THis way the purling will become flat after the dead loads are applied, and then only live load deflections will occur.
2. It will DEFINITELY NOT behave like an arch in any normal range of cambering (1"-3" over 25 ft, say).
I've been involved with light gage for awhile and I've never seen anyone camber floor or roof joists for the serviceability benefit. Not that it wouldn't help, but I don't think it would be economical and (nothing against drywallers) they may not be installed properly.
jetmaker-
Effectively, YES. There may be some residual stresses due to mechanically cambering a member, but these stresses are minimal, practically indeterminate, and have negligible, if any, effect on the design stresses. Also, as structuresguy stated, the precambering has no effect on amount of deflection.