Production is not the bottleneck, see 075. Once the design has been established, several ships could be built in quick succession.
But you will plan your shipyard capacities in such a way that ships always have to be built. When the last ship is launched, the subsequent new building must replace the first ship.
I don't see China building carriers simultaneously in multiple yards. Alternating to broaden the industrial base and test out the supply chain maybe. I simply don't see a requirement for a large carrier fleet especially if nuclear-based.
If China's eventual carrier fleet will be 8 and each has a 40 year service life, the steady state production rate would be one launch every 5 years (#17 was launched in 2017). At this rate, it is neither necessary or economical to build carriers in multiple yards. However, as in a poker game, if you bluff and raise, you just may force your enemy to fold first. That is a good-enough reason to build a second 003 in Dalian just to demonstrate industrial capabilities. A secondary reason is to use the 003 construction qualities to award the 004 contract to promote competition between Northern and Southern yards.
Longer term (2035), how can the U.S. maintain more than 8 big tops in the fleet with a declining budget? China has signaled a clear intention to scale up the submarine force both nuclear and AIP. This may explain in part of the U.S. interest in CVL's which are more defensive and perhaps more ASW-oriented. Counting CVL's as 'carriers' also conveniently circumvents the Congressional Mandate. With a continued buildup of China's DDG surface fleet, it may compel the USN to focus on the submarine fleet to assume much greater offensive roles (SSGN's) and reduce its surface fleet size. One also must question how much longer the U.S. can/want/able to act as the World Police given its ever stronger isolationist feelings.
2020 + 40 is already 2060 when China will be carbon-neutral and largely energy independent. Traditional maritime choke points won't matter much with possible Northern Passage and the overland BRI alliance with the Islamic Arc (Syria/Iraq/Iran/Afghanistan/Pakistan/Malaysia/Indonesia) fully developed. Perhaps the arms race then will be 90% cyber-warfare and combined-cycle this-and-that. How many carrier groups does China need to parade around the world?
In short, the current fascination with big flattops maybe as silly as the French building the Maginot Line or the Japanese completing the Yamato and Musashi. If so many people are questioning the utility and survive-ability of big carriers today, how about 10 years from now? All things considered, I don't see why China would be in a hurry to build up the quantity. Everyone everywhere already acknowledge China's ship building speed. Further demonstration is wasteful, unwarranted, and unnecessary.
#18 launch in 2022, #19 launch in 2027 (100th Anniversary of the PLA, J-XY in LRIP?), #20 launch in 2032-2034 (Name it after Xi's birth place BEIJING as a retirement gift) is a decent pace. There will always be new toys to spend USD$6B a pop on.