It seems to me that one of the more efficient means of interplanetary travel is one in which you don't have to bring the fuel mass or the power along with you from the Earth. Recent discoveries have shown large quantities of H20 (of what seems to be in some areas relatively accessible) in the polar regions of the Moon, some areas of which also offer the advantage of 24/7/365 sunlight.

The materials required to manufacture solar panels are present in the lunar regolith though it's unknown whether or not this is doable in the near term (but both the cost and efficiency ratios are improving year over year, making it more feasible to bring them with you if you can't manufacture them in-situ). The polar locations could potentially be excellent locations for very productive solar arrays to be installed, giving a potential MB access to much more significant amounts of power(which we can potentially do fun stuff with) than missions like Apollo had.

Many of the systems that were being designed for Constellation had a dual Mars-Moon role (such as the Orion vehicle), and there is some push in the Space community for a cycling vehicle of some sort (Buzz Aldrin), so why don't we marry all of these things together?

The family of missions that I would propose would be a clear and direct series of intermediary operations that would eventually allow us to get a manned mission to Mars surface, learning much as we "Island Hop" from LEO to the Moon, then on to Phobos, and on to Mars surface.

First, a cycling ship, capable of refueling in orbit and going from LEO to the Moon or Mars. Crews could be delivered to the ship in LEO on something like the SpaceX Dragon. I think having modular capabilities in the Cycling ship would be important, different vehicles could be attached to the crew module depending on mission parameters. A lander could be attached for surface missions to the Moon or Mars. I would send the first mission to the polar regions of the moon to survey for possible ice-mining operations.

If we find that it's feasible to mine the ice and manufacture fuel on the surface we could add extra capacity tanks onto the lander so that fuel could be sent up to the Cycling Ship in LLO. Power for turning the H20 into fuel could be provided by a solar array in 24/7/365 sunlight.

An Apollo 8 like mission to open up Mars could go to Phobos as an intermediary step, we could attach a Phobos docking module for the first trip instead of a lander. The cycling ship could then head back to Earth and pick up a lander while several astronauts left back on the Phobos module tele-operate rovers to scout for landing sites for the manned exploration of the surface, as well as survey Phobos for ice-mining potential. Once the Cycler returns with a lander we could then head down to the surface.

Down the road I could see us using some of the resources and knowledge we've gained in this first suite of missions developing more significant solar arrays in the lunar pole regions that could potentially power laser satellites in LLO that could be utilized for interplanetary flight operations.

You can argue over costs and feasibility of specifics (I don't think there are any ideas that are too out of this world), but the mission priorities would stress the reusability and modularity of hardware, promote utilization of in-situ resources, minimize launch weights from Earth and give NASA a clear and direct path where prior missions pave the way for future ones.

Now I just need a few hundred billion dollars, anybody have some spare change?