[Spaceships] How should a large station... duck?

[DataPacRat]

Say that a trans-to-post-human culture decides to implement one more kind of backup of itself, in the form of a complex ecology intended to last for ridiculously long periods with minimal nudging, inside a SM+34 McKendree Cylinder habitat. (Call it about a million square km of surface area, mass around 10 quadrillion tons; something vaguely like https://www.orionsarm.com/im_store/CYLINDER2.jpg , though with much larger solar collectors. And we can handwave construction methods with the magic words "von neumann".) And let's say that physics allows for spacecraft to accelerate to goodly fractions of lightspeed.

Without drifting too close or far from its home star, how can the cylinder's orbit be jogged, to make it more difficult for a hostile near-c interstellar impactor to target the cylinder?

Eg, can solar sails (or some other propulsion without reaction mass, such as a magsail) be arranged so the cylinder can 'tack' along its orbit, sometimes going a bit prograde and sometimes retrograde (or radial and antiradial, or normal and antinormal), so its average orbit remains at the right distance? Should I dive into the orbits of Janus and Epimethus, and have two cylinders who can "trade orbits"?

Would it help if the star this cylinder orbits was deep enough in a nebula to make it infeasible to pick out the cylinder from afar?

I'm hoping to find some numbers that add up to let the large cylinder be sufficiently unpredictable to dodge any good-sized near-c impactor before that impactor can change course. I'd rather stick with Reality Classic, but am willing to posit a reactionless drive (within certain limits). If that doesn't work out, I'm willing to shrink the cylinder to the point where such manoeuvres become possible, though I'd still like it to be large enough to be a BDO with lots of interior space to be explored. If /that/ doesn't work out, I'm willing to consider alternate approaches, such as Large Numbers of smaller habitats (eg, https://orionsarm.com/im_store/JenkinsDysonSwarm.jpg ), though I'd have to significantly re-jigger my intended narrative, so it's a last resort.

How many tricks am I missing?


-----8<-----

Preliminary build:

* SM +34, USL. TL10, no superscience
* SM +34, USL. TL10, no superscience
- 10 Quadrillion tons, length 1M yards (cylinder: length 914 km, circumference 1106 km, radius 176 km, diameter 352 km), dST/HP 150M, Hnd -10/5
* Design Features:
- Spin Gravity (0.2G, $100 Quintillion, -2 to handling)
- Self Healing (TL10): heals 1.5M HP/day, $200Quintillion
- Stealth Hull (-8 to detect): $1Quadrillion
- Chameleon Hull (-4 to see): $700T
* Design Switches:
- Exposed Radiators
- Slower Industrial Systems
- Pyramid 34: Armor Volume (9 armor systems: dDR * 1.4)
* Systems (18 + 2 core):
8 * Armor, Stone: dDR 30,000 (30k * 8 * 1.4 = dDR 336,000, semi-ablative: every 10 points of damage removes 1 point of DR)
1 * Armor, Nanocomposite: 500k dDR, $50Qn (500k * 1.8 = 700,000, total dDR 1,036,000, avg dDR 345,333 each for fore/central/aft)
10 * Open Space: each 500M areas (each 25M acres, total 250M acres, max of ~200M humans using traditional farming), Total Automation (replace 300M workspaces, $1.5Quintillion), $200T
1 * Smaller systems: 30 SM+31 systems (or even more smaller ones):
- SM+31 Cargo Hold: 15Qd tons
- SM+31 Control Room: C20, Comm/sensor 30, 15360 control stations, Total Automation (replace 30B workspaces with $150Qd), $600Qd. -1 to Handling/Stability.
- SM+31 Enhanced Sensor Array: Array Level 32, Total Automation (replace 30B workspaces with $300Qd), $600Qd
- SM+25 Robofac or Vatfac: $15T/day, Total Automation (replace 30M workspaces with $150T), $30Qd
- SM+25 Habitat: 2B cabins, Total Automation (replace 30M technicians with $150T), $300T
- SM+20 Hangar Bay (for robotic miner ships): capacity 1B tons (max SM+16, door diameter 700 yards), launch 3125k tons/minute, Total Automation (replace 3M workspaces with $15T), $3T
- SM+25 Mining: 1.5B tons/day, Total Automation (replace 30M workspaces with $150T), $300T
- SM+25 Refinery: 5B tons/day, Total Automation (replace 30M workspaces with $150T), $300T
- SM+25 Power Plant, Reactor, Fusion: 1PP, 400 years, Total Automation (replace 30M workspaces with $150T), $1.5Qd (refuel $150T/400y, ~= $375B/y)
- SM+25 Solar Power Array: 1PP, $1.5Qd
- SM+19 Robot Arm: Total Automation (replace 30k workspaces with $150B), $3T
- SM+18 Weapons battery, Hidden: 1B * SM+9 Weapons, Major Battery (TL10 UV Laser: 4dx5 damage, range L, fixed mount). $22.5Qd. (333M can be aimed at any one target)
- SM+14 Weapons Battery, Hidden, $22.5T: 1B * Missile launcher, 20cm: sAcc -2, range L, rcl 1: SM+5, 7 shots ea
- Ammunition, 1T shots: 200M tons, $200T
- Ammunition, 1T shots, 100 kt nuclear: 200M tons, $300T

* Price so far: $366.71 Quintillion.

----->8-----