Delvo would be an excellent product for this situation (and no I don't work for deGussa). I would not recommend a normal retarder unless it is monitored like a hawk through production as cement chemistry does and will change thus influencing the effectiveness of a normal retarder, not to mention the varying external temperatures you'll be seeing.
Also, as you are air-entraining this mixture, you'll want to monitor the air content, preferably in a hardened state not just with an air meter. Air-entraining agents stabilize the bubbles that develop when mixing. The air content will fall with time and the trend will be towards a coarser bubble structure. Redosing on-site is not always effective and an air meter may not capture this. This will make the concrete freeze-thaw susceptibile and reduce the lifespan of the structure being constructed. Perform a test batch at the ready-mix plant, cast a cylinder after initial mixing is done, set the ready-mix truck aside for 2 hours or so and cast another sample. Have the cylinders tested for hardened air void parameters per ASTM C 457. I would also throw in a couple of production checks as well through the project. The test is not easily done and can be expensive. Given that the nearest plant isn't close, I assume the nearest concrete petrographer is probably a lot further a way. God Bless FedEx though...
I would be deeply concerned about mixing the rock, sand, cement, and admixture and then adding the water on-site. Admixtures can be sucked into dry aggregates rendering them useless. The last thing you need is to transport 10 CY of concrete 2 hours and then find out you can't get the air content in spec. (I might add, at the time you find out it isn't working, you could already have 100 CY in transit).
