FaultBench-Industrial Research documentation
Research documentation ยท Experimental record
Agent dropout / Transient

Fault-tolerant-trained MARL

Transient agent dropout evaluation

Fault: Agent dropout
Duration: Transient
Recovery mechanism: Fault-tolerant-trained MARL

Experiment definition

The experiment evaluates Fault-tolerant-trained MARL under a transient agent dropout injection. Fault onset is randomized between environment steps 40 and 60.

Expected evaluation record

MetricRecorded value
Degradation-area30.49 ± 5.57 (corrected; mean ± 95% CI, n=72)
Safety-violation excess187.7 ± 35.2 (weighted, control-subtracted; mean ± 95% CI, n=72)
Fraction of time healthy0.521 ± 0.033 (mean ± 95% CI, n=72)
Relapse count0.68 ± 0.17 (mean ± 95% CI, n=72)
False-recovery raten/a - no fault detector (trained policy only)
Intervention recalln/a - no fault detector (trained policy only)

Values are means with 95% confidence intervals across every target × duration-fixed combination shown on this page (see graphs below), 8 evaluation seeds each, for each of 3 training seeds. See the Results page for the cross-mechanism comparison.

Experimental trajectory graphs

One representative run (evaluation seed 0 of 8, training seed 1 of 3) per target combination from the full sweep. See the Results page for the multi-seed aggregate comparison across mechanisms. The graphs for every seed are archived at doi:10.5281/zenodo.23304785 (files available from 10 November 2026).

Sorting Agent Target

Agent dropout transient fault, Fault-tolerant-trained MARL, Sorting Agent Target
Fault-tolerant-trained MARL - Sorting Agent Target - evaluation trajectory under transient agent dropout (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep, training seed 1 of 3.

Pressing Agent Target

Agent dropout transient fault, Fault-tolerant-trained MARL, Pressing Agent Target
Fault-tolerant-trained MARL - Pressing Agent Target - evaluation trajectory under transient agent dropout (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep, training seed 1 of 3.

Both Agents Target

Agent dropout transient fault, Fault-tolerant-trained MARL, Both Agents Target
Fault-tolerant-trained MARL - Both Agents Target - evaluation trajectory under transient agent dropout (fault onset randomized between steps 40-60), seed 0 of the 8-seed sweep, training seed 1 of 3.

Notes

This page is intentionally separated from the paper-facing results so that full-resolution plots, additional seeds, and diagnostic figures can be documented without overloading the publication.