If this is a suspended-tube radiant you'll get some benefit out of a radiant barrier, but if it's plated with aluminum heat spreaders in full contact with the sub-floor the benefit is almost vanishingly small. The emissivity of aluminum is EXTREMELY low, and the fraction of radiated heat transfer is miniscule- if there's (nearly) nothing to reflect, the radiant barrier has no function- it'll work about as well as a sheet of poly (or bubblepack). Claims of "R20 equivlent" are completely bogus, based on a totally different ASTM test configuration that is not relevant here.
The emissivity of the bare wood is much higher, but it's mostly covered with low-e aluminum plates, and even if the heat was all above the floor, the most you'd get out of RB is a ~25% (when very new & shiny) reduction in downward heat flux, since most of the heat loss would still be convected & conducted, not radiated.
Plated or suspended, you need some sort of insulation that has an R value determined by ASTM C518, a conducted & convected heat transfer block. In hot-side up configurations even low-density unfaced fiberglass batts still do OK, since with the warm side up the induced convection is still quite low. If above a full basement that's reasonably air sealed it doesn't have to be perfect either. It's cheap and (in this case) effective. (DO foam seal the band joist & sill ends of the joist bays, to prevent cross-drafts throuh/above the insulation though.)
Above conditioned space R11-R15 is fine, but if it's above a cold drafty crawlspace/basemen R19 min. If plated, snug it right up against the plates (but with minimal compression.) If suspended, cut strips of radiant barrier (bubble or aluminized polyester, whatever is cheapest) that fits loosely, resting atop the fiberglass, but hang the insulation as low on the joist as possible, with the largest possible air gap between the RB & tubing. You could cut the RB wide and staple it, but that just adds a lot of labor with minimal benefit.
Rigid board insulation is about the most expensive way to go, and may be very difficult to install around obstacles. In some installations it's easier (but not usually cheaper) to spray 4-6" of half-pound foam (Icynene, et al) directly to the underside, (but woe to whomever has to repair a leak from an ill-placed flooring nail if you go that route. If you DO go with rigid board stuff, EPS (beadboard) is usually the cheapest per unit R value, and you can get it with single side foil-facer (which is an effective radiant barrier for suspended tube stuff.) You'd need a minimum of 2.5" of EPS or 2" of XPS (pink/blue board) or 1.6" of polyisocyanurate over a decently air sealed basement. Then, in order to meet fire code you'd be required to put a half inch of gypsum between it and the basement (a ceiling) as a thermal barrier to keep the foam from lighting up/creating toxic smoke in a fire situation. (With fiberglass there is no such requirement.)
Use UNfaced fiberglass batting and PERFORATED radiant barriers, or you run the risk of creating seasonal mold conditions. You want to avoid placing a vapor retarder between the basement and first floor if you don't have to, or else control the humidity in both the basement & first floor to guarantee that the average relative humidity anywhere on the joists stays below 70%. (In summertime 60% RH air at 75F on the first floor exceeds 70% at the point in the insulation where the temp drops below 70F. If there is a vapor retarder anywhere on the basement side of that depth, you have mold-enhancement going on even if you're dehumidifying the basement, since it's on the first-floor side of the hygric zone created by the vapor retarder.