More than ten years ago there were plans to build large antenna arrays for DSN, about 400 antennas of 12m diameter.
What happened to these plans?
I will leave room for another answer to discuss the specifics of the current status of plans to build arrays of microwave dishes for the Deep Space Network.
There is one practical problem with arrayed receivers that I read in a NASA document somewhere, perhaps a DESCANSO report or elsewhere, that talks about limitations of beam power in beams formed by arrays, rather than a single aperture. The problem had a specific recognized name that I can't recall, but the idea is that arrays have a lot of empty space between dishes, and that will always lead to significant power loss to a giant cloud of side beams when you do the calculation quantitatively.
These side beams also pick up noise in receive mode, so if a planet or the Sun is nearby, there are situations where a sparsely sampled pattern will be far noisier than a solid continuous dish for practical reasons. I still need to find references on this one though.
I will also point out also that Ka-band is the highest practical atmospheric "water window" before optical†, as discussed in @BobJacobsen's thorough answer to the question What is the highest non-optical frequency used or tested for use in deep space communication?
†optical can include infrared bands as well, the distinction between microwaves and optical is not fundamental and gets blurred more as THz technology advances. For the purposes of this answer let's say "optical" refers to wavelengths of roughly one or two microns and shorter, which includes most well-developed compact laser light sources that you'd like to put in space and depend upon. There are cascade quantum well lasers for THz, but these aren't really "there" yet in terms of putting on spacecraft.
Optical is really the next step, done either from the ground, or because the cost of building and launching to LEO is plummeting, done from transponders in LEO that can then link optically or via microwaves to very modest ground stations.
technology moves on and displaces great things with new things that will become great as well.
In this particular case it will be optical communications that will displace the idea of a major upgrade of DSN to dish arrays for two reasons:
Behold the "Space Communication and Navigation Testbed" notional architecture found in Glenn Research Center's Richard Reinhart's slides Space Communication and Navigation TestbedCommunications Technology for Exploration for the 2013 ISS Research and Development Conference:
One of your sources: The DSN Array Development Program, Sander Weinreb, 818-354-4065 May 23, 2002 shows a plot with optical at a data rate much lower than several radio bands in 2002 with data points, and an estimated order of magnitude jump predicted 8 years later in 2010 (no data points of course, just a displaced curve).
We're now almost in 2020 and the technology has been demonstrated several times. The technology works nicely and is extensible.
As another of your sources (A Cost-Effective Array-based DSN (DSAN); J. Statman (with contributions from B. Geldzahler, L. Deutsch, and R. Preston) 7/29/2004) mentions in 2004:
– Optical communication, which will take at least another decade to mature and two to be operational, has development risk and may not be appropriate for all missions or mission phases
We're 1.5 decades past that and things are progressing nicely. Whatever "may not be appropriate for all missions or mission phases" refers to, it probably won't warrant building DSN arrays.