No clear winner. The survival gap is under 10 percentage points and the balanced-score gap is under 0.05, so neither SolvSRK nor the best baseline clears the win threshold. Either works — choose on cost, licensing, or integration effort. All verdicts →
20 targets: 8 maneuvering (omega 0.1-0.5 rad/s), 6 cruise (omega 0.01-0.05 rad/s), 6 hovering (omega=0). Coordinated-turn dynamics with acceleration decay. Designed for SF-3 high-dimensional regime: UKF requires 281 sigma points per update at this dimension, EKF needs only 1 prediction + Jacobian.
Bar-Shalom et al. (2001) Ch. 11; Blackman & Popoli (1999) Ch. 7-8
Canonical RHS excerpt from the registered callable used for this benchmark cell. Expand it to verify the state equations; it is not a standalone runnable fixture.
def rhs(t, y):
d = np.empty(dim)
for i in range(n_targets):
base = 7 * i
x = y[base]
yy = y[base + 1]
z = y[base + 2]
vx = y[base + 3]
vy = y[base + 4]
vz = y[base + 5]
a_m = y[base + 6]
v_mag = np.sqrt(vx * vx + vy * vy + vz * vz)
v_safe = max(v_mag, _VEL_FLOOR)
omega = omegas_arr[i]
d[base] = vx
d[base + 1] = vy
d[base + 2] = vz
d[base + 3] = -omega * vy + a_m * vx / v_safe
d[base + 4] = omega * vx + a_m * vy / v_safe
d[base + 5] = a_m * vz / v_safe
d[base + 6] = -a_m * inv_tau_a
return dCanonical RHS excerpt captured from the same registered callable used for the published benchmark. Frozen closure values are summarized below; helper imports and solver settings are intentionally omitted.
Fingerprint
Spread: extreme
Default noise: low
Recommendation snapshot
Clean best: SolvSRK
Noisy best: SciPy BDF
Coverage
14 solver arms · clean + 5 noise levels
Ranked on survival, precision, and speed
20 seeds/cell default · 14 arms · TRL 4–5 · simulation-lab validated · this page: 20-Target Large Swarm Tracking (20-target-large-swarm-tracking)
Governed SolvTune benchmark freeze; per-arm medians only. RHS definitions and raw trial rows are not published.
Self-reported by Resonix Labs · not independently verified
Pick an objective and a noise level to rank all arms on survival, median SCD, median nfev, and median wall time. Medians across seeds.
Objective
Best overall trade-off of survival, precision, and speed.
Noise level
| # | Solver | Survival | SCD | nfev | Wall | Score |
|---|---|---|---|---|---|---|
| 1 | SolvSRK | 100% | 10.3 | 8,972 | 699 ms | 0.864 |
| 2 | SciPy RadauSciPy | 100% | 10.2 | 5,154 | 325 ms | 0.862 |
| 3 | Tsit5external | 100% | 8.3 | 1,620 | 933 ms | 0.818 |
| 4 | SciPy LSODASciPy | 100% | 8.3 | 1,033 | 43 ms | 0.816 |
| 5 | SciPy DOP853SciPy | 100% | 8.0 | 518 | 23 ms | 0.810 |
| 6 | SciPy RK45SciPy | 100% | 7.9 | 1,658 | 73 ms | 0.806 |
| 7 | CVODE Adamsexternal | 100% | 7.4 | 974 | 51 ms | 0.795 |
| 8 | SciPy RK23SciPy | 100% | 7.3 | 10,817 | 491 ms | 0.792 |
| 9 | CVODE BDFexternal | 100% | 6.9 | 1,997 | 96 ms | 0.784 |
| 10 | SciPy BDFSciPy | 100% | 6.5 | 1,292 | 101 ms | 0.772 |
At Clean, best balanced arm is SolvSRK.
Values are medians across seeds, measured by Resonix Labs on Resonix hardware and not independently verified; nfev and wall are on reference lab hardware (indicative). Under injected noise only SolvSRK and the SciPy arms are run. How we measure accuracy → · Verification status →
SolvScout · free
This page shows one published benchmark cell. SolvScout fingerprints your ODE, compares it to the full corpus, and recommends a solver with the same survival / precision / speed ranking you see here — including when a SciPy arm wins.
SolvSRK · 30-day trial
SolvSRK is the stiffness-adaptive integrator behind the SolvSRK column in these tables. Create an account, activate a machine, and take a 30-day trial — same binary you'd ship after purchase.
Replace the access date. Pin the freeze ID and library versions when comparing against a later export. Cite it as what it is — a self-reported vendor benchmark, not an independently verified result. The note field says so; please keep it.
@misc{resonix_evidence_20_target_large_swarm_tracking_2026,
title = {Resonix Evidence Portal: 20-Target Large Swarm Tracking},
author = {{Resonix Labs (Canada) Inc.}},
year = {2026},
howpublished = {\url{https://resonix.tech/evidence/problems/20-target-large-swarm-tracking}},
note = {Self-reported vendor benchmark; internally generated by Resonix Labs and not independently verified. Accessed YYYY-MM-DD. Freeze 2026-08-13; libsolvsrk 2.3.0; SciPy 1.14.}
}TRL 4–5 · simulation-lab validated · 398 problems · 14 solver arms · clean + 5 noise levels
Freeze: 2026-08-13 · scipy 1.14 · libsolvsrk 2.3.0 · Methodology
Self-reported by Resonix Labs · not independently verified · Verification status