In this project, the approach is taken that the first missions to settle
the Moon will focus on developing the robotic infrastructure needed to
do construction entirely without humans present. Such robots do not yet
exist, but the field is advancing quickly. Development of that
technology has a wealth of applications here on Earth, it would make
sense to make such R&D a central part of a settlement space program.
A few missions
would be needed where a minimal crew go to support the prototype robots,
testing and fixing them, until there is confidence the robots can be
operated remotely from Earth. After that, the robots build the
first permanent habitat and its infrastructure without
anyone on-site. Once that hab is near completion,
again a crew
arrives. They test everything, maybe make some minor final touches.
Then they stay.
After that the human population ramps up quickly and they operate the
robotic machinery from within the habs.
So, great things come from great robots. Rovers in space are already designed to handle factors not present on Earth. Rovers and robots on the Moon will have extra challenges that have not yet been solved. The list below takes them in ascending order of difficulty:
The videos in the sidebar show we've made big advances on these things, but there is a ways to go. It is hard to overstate how much of a difference robots with these capabilities make. They are absolutely necessary for success.
If there was a way to build a habitable shelter on the
Moon that wasn't super hard, we'd already be there. It
is important not to underestimate how hard it is to do
this. But, the needed technology is coming together and
turns out to change the equation of what the best method is.
Most designs for lunar construction up to now have focused on batch processes that don't require much finesse (such as making lunar concrete blocks). Few designs have contemplated use of in situ materials, though, because of the difficulty of delivering the machinery to work them, and the probability such machinery wouldn't last long in such a harsh environment. Almost all serious designs send finished modules built on Earth.
If you plan to stay, you have to be able to build there. If you can build there, you have something you can sell. If you are good at it, you can build for space, too, and that's our whole idea.
All construction designs here are centered around the use of melted regolith, in particular the basalt of the lunar maria. When basalt is drawn into fibers it has good tension strength. A material that is strong in tension is the key to space development. Holding in an atmosphere requires it. Basalt rope and cloth can be made with no additives other than a thin application of a sizing chemical to the surface of the threads. This makes construction with basalt much easier.
Some highlands materials spun into fibers will also have good tension strength, but likely need to be processed to change their chemical composition before that is true, and such processing would take considerable infrastructure. Once they are, they might also need to be embedded in a matrix, like fiberglass is embedded in plastics on Earth. Glass fibers aren't used for rope or fabric on Earth because it only takes slight damage to cause the fibers to lose their strength. On the other hand, quite possibly the different chemistry and environment of the Moon means that they would work well without needing a protective matrix. Still, unless deposits very high in silica are found, resource-intensive processing will be needed in the highlands to produce materials capable of containing an atmosphere. That complicates construction of habitats a great deal, a principal reason why only a small human colony will be made at the pole, and then construction will be concentrated at Lalande.
All these designs absolutely depend on the use of agile robots. They don't have to be carpenters, but they do need the manual skills of your typical 5 year old. Guided by human operators on Earth, that would be enough. Several development missions would need to go to the Moon to test and refine that technology.
If you have those robots, what follows is the best way to build. By far, no contest. It is far less energy intensive, and far more adaptable, than anything else I've seen. The equipment needed to start is far less, and less complex. Except for the robots, of course, which are so broadly useful they are still quite mass-efficient, even if they need a bunch of modular drop-in replacement parts shipped with them.
So, if you perhaps calculate that such robots are a long way off, still by the time Cernan's Promise was fully developed, they would exist and have been used for the bulk of that construction. We can quibble later about the most plausible way to present the first few missions, when those robots are in development and can't be relied on. The winds blow very strongly towards their development, so it is sound policy to explore what they can do.