Some speculation that the Earth doesn't have Venus's atmosphere due to the presence of the moon. So given a moon to Venus may be sufficient to thin out the atmosphere over sufficient time. Put Ceres in a 90 minute orbit might do the trick and spin Venus up at the same time. (Wait. Roche limit. Hmm. Another ringed planet. Would a band of rings have a net cooling effect?)
A more massive Mars could hold at atmosphere for more time. A molten core may provide a magnetic field to keep the solar wind from splitting water.
Change the orbital distances with some caution, and a lot of simulation time. Orbital dynamics is tricky. Resonances where the period of one is a small integer ratio with the period of another are particularly tricky.
If you accept a solar system formed with massive terraforming by a previous civilization, or working on a very long time scale, bombarding Mars with enough comets would establish at least a transient (few tens of millions of years) atmosphere. Use something like Ceres, and try for a gentle collision and you would get a mass boost, plus a molten core again. Drag one of the ice moons in and siphon off enough water to fill martian seas.
Venus could be uninhabitable if you can get rid of most of the atmosphere (matter transport to mars...) and paint most of it titanium dioxide white. You want a material that is black in the infrared, and white in the visible to minimize solar absorption, and maximize radiation. Even with this I think you would have a wide equatorial band that was too hot. While your at it hit with enough rocks to give it a decent spin.
Remember also that you have 6 other potential spots in the trojan positions 60 degrees ahead and behind these 3 planets. Could be good parking places for ice moons to thaw, or to have industrial worlds. The energy to move from one trojan point to the other in the same orbit is a matter of how long you are willing to wait.
There are other Lagrange points:
* L3 "Counter earth" on the opposite side of the sun:
* L1 "Sub solar" where it between earth and sun, with the excess gravitational attraction of the sun balanced by the gravitation of the earth.
* L2 "Supra solar" where it is just beyond the earth along the earth sun line.
None of these are long term stable. The counter earth I think is close to neutral stability. L2 and L1 are unstable, and perturbations grow rapidly. If you have the resources to move planets around, however station keeping should not be a big issue.
If you need more real estate, play with orbital dynamics and sharp angles to the ecliptic. These have less coupling, and so you should be able to orbit a planet at right angles to Earth's orbit between Earth and Mars. Note that the energy to reach these orbits is horrendous.
Another solution that is elegant, but difficult is a klemperer rosette. N bodies in an ellipse co-rotating about their mutual center of gravity. This has been used in two stories that I know of: The Fleet of Worlds of the Puppeteers of Niven's known space universe, and in Arthur C. Clarke's novel either Against the Fall of Night or The City and the Stars, where reference is made "It is lovely to watch the coloured shadows on the planets of eternal light."
AFter this, it's time to look at building ringworlds.