I don't think that's necessary. The loss of atmosphere by Mars may have been significant on the timescale of its five-billion-years history, but not on the timescale of a human settlement project.

The mass of the Martian atmosphere is about 25 teratonnes [1]. According to the measurements by MAVEN, the rate of mass loss by Martian atmosphere due to solar wind is 100 g/s [2]. At this rate it would take

    25e12 * 1000 * 1000 * 0.01 / 100 / 3600 / 24 / 365.25 = 79 million years

for Mars to loose 1% of the mass of its atmosphere. Even if the rate of mass loss were 10 times higher than measured by MAVEN (it's known to increase during solar storms and in perihelium), it'd still take nearly 8 million years.

Looks like if we can increase density and pressure of Martian atmosphere as GP suggested, we don't really need to worry about the lack of magnetospheric protection.

[1] https://en.wikipedia.org/wiki/Atmosphere_of_Mars [2] https://www.nasa.gov/press-release/nasa-mission-reveals-spee...

I think it is still an open (and hotly debated) question on if planets need magnetospheres to retain an atmosphere.

Mars has isolated pockets of "mini magnetospheres" that are suitable enough to block charged particles.

http://www.space.dtu.dk/english/Research/Universe_and_Solar_...

Ah yeah, forgot about that. And restarting the core dynamo seems pretty infeasible regardless or our technological capabilities.