Overview Evidence tour Methodology Dead zones Problems

11-compartment PBPK model

ADVANTAGES2 · dim 11

SolvSRK wins. At the comparison noise level, SolvSRK beats the best baseline by at least 10 percentage points of survival, or by at least 0.05 balanced score when survival is tied. Use SolvSRK for this class of problem. All verdicts →

Full physiologically-based PK model (arterial/venous/lung/liver/kidney/gut/muscle/fat/brain/skin/RP); S2 from fast hepatic vs. slow fat redistribution

Biomedical & bioprocess

Problem definition

Canonical benchmark implementation

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.

Show canonical RHS excerpt

def rhs(t: float, y: np.ndarray) -> np.ndarray:
    # y = amounts [mg] in each compartment
    C = np.maximum(y, 0.0) / V          # concentrations [mg/L]
    C_art = C[0]                         # arterial concentration
    dy = np.zeros(11)
    # Arterial pool receives lung venous output, pumps to tissues
    Q_co = Q[2]   # cardiac output
    # Lung compartment (mixes venous pool)
    C_ven = C[1]
    dy[2] = Q_co * (C_ven - C[2] / Kp[2])
    # Liver: portal + arterial input, intrinsic clearance
    Q_hep = Q[3] + Q[5]   # hepatic blood flow (portal + gut)
    dy[3] = Q[3] * C_art + Q[5] * C[5] / Kp[5] - Q_hep * C[3] / Kp[3] - CL_int * fuP * C[3] / Kp[3]
    # Kidney: arterial input, renal clearance
    dy[4] = Q[4] * (C_art - C[4] / Kp[4]) - CL_ren * fuP * C[4] / Kp[4]
    # Gut: arterial input → into portal
    dy[5] = Q[5] * (C_art - C[5] / Kp[5])
    # Remaining perfused tissues (muscle, fat, brain, skin, RP)
    for i in [6, 7, 8, 9, 10]:
        dy[i] = Q[i] * (C_art - C[i] / Kp[i])
    # Venous return (sum of tissue outflows)
    Q_ven_return = sum(Q[i] * C[i] / Kp[i] for i in [3, 4, 6, 7, 8, 9, 10])
    dy[1] = Q_ven_return - Q_co * C_ven
    # Arterial: lung output
    dy[0] = Q_co * (C[2] / Kp[2] - C_art)
    return dy

Parameters: CL_int = 30
CL_ren = 10
Kp = [1, 1, 2, 8, 4, 6, 3, 25, 2, 4, 5]
Q = [0, 0, 390, 81, 60, 52, 101, 14, 43, 18, 21]
V = [1.65, 3.3, 0.5, 1.8, 0.31, 1.1, 28, 14.5, 1.45, 3.5, 1]
fuP = 0.05
Initial condition: y(0) = [100, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
Horizon: t ∈ [0, 72]

Canonical 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: high

Default noise: none

Recommendation snapshot

Clean best: SolvSRK

Noisy best: SolvSRK

Coverage

14 solver arms · clean + 5 noise levels

Ranked on survival, precision, and speed

Versions & freeze

Methodology →

Freeze: 2026-08-13
libsolvsrk: 2.3.0
SciPy: 1.14
SUNDIALS: CVODE (bundled backend)

20 seeds/cell default · 14 arms · TRL 4–5 · simulation-lab validated · this page: 11-compartment PBPK model (11-compartment-pbpk-model)

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

Results matrix

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%	13.1	100,416	1.11 s	0.932
2	SciPy RadauSciPy	100%	10.6	4,194	113 ms	0.871
3	SciPy RK45SciPy	100%	8.6	60,914	821 ms	0.823
4	SciPy DOP853SciPy	100%	8.1	54,086	695 ms	0.811
5	Tsit5external	100%	7.9	62,610	6.68 s	0.807
6	SciPy RK23SciPy	100%	7.8	37,196	568 ms	0.804
7	SciPy LSODASciPy	100%	7.3	1,920	21 ms	0.794
8	SciPy BDFSciPy	100%	7.0	1,385	55 ms	0.787
9	CVODE BDFexternal	100%	6.7	752	16 ms	0.779
10	CVODE Adamsexternal	100%	6.3	3,929	58 ms	0.769

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 →

Cite this page

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_11_compartment_pbpk_model_2026,
  title        = {Resonix Evidence Portal: 11-compartment PBPK model},
  author       = {{Resonix Labs (Canada) Inc.}},
  year         = {2026},
  howpublished = {\url{https://resonixusa.com/evidence/problems/11-compartment-pbpk-model}},
  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

Problem definition

Versions & freeze

Results matrix

Cite this page

Related