Live Multi-Scale Many-Body Gravity (N-body) Simulation

When galaxy N-body is enabled, gravitational dynamics will run across the entire generated galaxy — up to 100,000 star systems simultaneously. Systems will not follow pre-computed orbits or visual approximations; each one will carry a live state vector advanced by a gravitational integrator running in real time on the user’s machine.

The intended result is physically motivated emergent behaviour: star systems drifting through the disc, clustering into transient associations under mutual attraction, and responding to the mass distribution of the whole galaxy. High-density regions — the bulge and inner spiral arms — are expected to produce tighter clustering and more dynamic encounters than the outer disc. The large-scale structure of the galaxy would evolve visibly under the force law that generated it.

The gravitational algorithm planned for runtime is the same Barnes-Hut tree that drives the generation pipeline’s N-body relaxation pass.¹There is no plan for a separate “viewer” approximation — the runtime simulation and the simulation that produced the galaxy’s initial conditions are designed to share a common physical foundation.

While the galaxy integrates, the plan is for the user to enter any individual star system and run full N-body integration on its own bodies simultaneously. The local integration context is designed to be independent: it would maintain its own set of body state vectors and its own adaptive step size, tuned to the much shorter timescales relevant to planetary motion rather than galactic drift.

Both integration contexts are intended to run at the same time. The host star would advance through the galaxy under galactic-scale forces while its planets advance around it under local-system forces. The user would see both: entering a system during a live galaxy simulation would not pause or decouple the galaxy — the star the user is standing next to would continue to move.

This means a user observing a close galactic encounter between two systems would be able to zoom into one of them mid-encounter and watch how the perturbation propagates into the planetary orbits in real time. The local system response is designed not to be a post-processed result — it would reflect the actual gravitational perturbation being computed at that moment.

A single timescale slider governs both integration contexts simultaneously. Increasing the timescale accelerates both the galaxy simulation and the local-system simulation by the same factor. The two scales remain physically consistent: galactic position is updated from the same simulation clock that advances local body positions.

At low timescales, the galaxy moves imperceptibly while planets complete visible fractions of their orbits — this is physically accurate, since a planetary year is a negligible interval on the galactic timescale. Increasing the timescale compresses both simultaneously: the host star begins to trace a visible arc through the galaxy while planets complete many orbits per second.

The slider does not switch between two independent simulations running at different rates — it scales a shared clock. This means there is no drift between the two contexts over long runs: the star’s galactic position at any moment corresponds exactly to the planet positions at that same simulated time.