TRACE SUPPLY / INTERNAL OPERATIONS + LOGISTICS

One disruption.
A whole chain.
A decision to explain.

We study how a deviation connects assets, maintenance, spare parts and suppliers. TRACE Supply combines hypotheses about origin with consequence comparison to support a supervised decision.

Technical foundation developed · Synthetic studies closedIndustrial usefulness and each integration require assessment with agreed data and criteria.

TRACE SUPPLYRELATIONSHIPS + TIME
Asset
Spare part
Work order
ORIGIN → MINA

Hypotheses · Consequences · Alternatives

DECISION CONTROLResolve, request data
or abstain.
Technical foundation diagram. A person reviews and authorises any real action.

YOUR TEAM’S WORK

Two pathways.
One shared foundation.

Two starting points for continuity: what happens inside your operation, and how materials and products arrive or move to their destinations. They share dependencies, inventory and spare parts. The case is scoped around the decision you need to make and the available data.

↳

Joint development for both pathways. Adapters, data contracts and demonstrators with agreed contributions and criteria.

Explore R&D ↗
Conceptual scene of a spare-parts storeroom and industrial maintenance area

FROM SPARE PART TO OPERATION

The stock that protects.
The dependency that stops work.

CONCEPTUAL IMAGE · NOT A CUSTOMER INSTALLATION

AN EXAMPLE TO EXPLAIN IT

A delay does not always
stop operations.

Autonomy and an alternative can absorb a disruption. Explore this explanatory model to see why time and missing information change the decision.

A spare part pending for a maintenance task.

OUR OWN ILLUSTRATIVE MODEL

  1. Supplier
  2. Spare part
  3. Maintenance
  4. Operation

Simplified explanatory simulation. It does not run ORIGIN or MINA or reproduce a blind case. Uncovered days = max(0, effective arrival − autonomy). It does not calculate financial losses or model a complete operation.

Read the model and its limits ↓

WHAT WE HAVE DEVELOPED

Find the origin.
Compare the consequence.

Supply is a separate line. ORIGIN, MINA and ambiguity control have preserved versions and studies. Their synthetic results do not establish universal industrial applicability.

01 / TRACE ORIGIN

From partial signals
to origin hypotheses.

Ranks candidate causes and their temporal subgraphs. The V0.2 policy retains alternatives when confusion can change the consequence.

Inputs and scope +

Signals available at decision time, dependencies and profile restrictions. The internal score is not a calibrated probability.

02 / TRACE MINA

From a hypothesis
to its consequences.

Studies affected entities, absorption, first impact, recovery, obligations and economics within the model. Compares consequences and possible responses.

Inputs and scope +

Topology, capacity, inventory, consumption, lead times, buffers and obligations. The V5.2/R15 foundation is frozen as a candidate for controlled industrial evaluation.

V0.2 control: resolve, request data or abstain.

If two plausible causes imply incompatible responses, the policy requests a distinguishing datum available at T0 or retains abstention. A person retains authorisation of any real action.

HOW FAR WE HAVE REACHED

Results with context.
Visible limits.

The ORIGIN V0.2 closure on 21 August 2026 passed the frozen gates of an internal synthetic blind cohort. These figures belong to that study; external industrial validation is still pending.

400synthetic blind cases
207/240material cases resolved
33/240material cases still unresolved
0/80false activations on absorptions
01 / MINA V5.2-R15Consequences and absorption+

300 new synthetic cases and a separate supplement of 40 negative controls. Operational F1 0.9987; contractual F1 0.9813; economic WAPE 31.21%. The supplement recorded 0/40 false alarms. Economics remains a model estimate, not a validated industrial budget.

02 / E2E V0.1 → STRESSThe failure that changed the decision+

The first integration was studied on 400 cases. A later 320-case discriminative stress failed four performance gates: choosing a nearby cause could materially change the response. That FAIL is retained and motivated V0.2 control; its metrics are not mixed with the later study.

03 / ORIGIN V0.2Denominators and ambiguity+

The reference cause remained in Top-2 in 400/400 cases. All 64/64 material Top-1 errors were detected and 0/64 were accepted as safe decisions. Of 240 material cases, 79 resolved initially and 128 after a T0 datum request; 33 remained unresolved. Consequence agreement on the 207 resolved cases is measured against MINA with the hidden cause, not against a real company.

04 / INDUSTRIAL TRANSFERWhat is prepared and what is pending+

A file-based exchange protocol and logistics compatibility work exist: provisional mapping, schema and validator. The final contract and external blind cohort remain open. Production installation, a cloud service, ERP/TMS connections and a public installer require their own development and agreement.

05 / DOCUMENTED LIMITSReproducibility and current use+

The V0.2 closure preserves aggregates and hashes, but not a historical per-case ledger that can be reopened. Four historical regression tests could not run because dependencies or corpora were missing and are not counted as passes. The next evaluation must preserve its per-case record from the start and measure usefulness against the agreed procedure.

Technical status: synthetic blind ambiguity-control PASS. Industrial accuracy, financial savings and autonomous operation are not established by these studies.

R&D PARTNERS / TRACE SUPPLY

One industrial use case.
Joint development.

We are looking for companies to develop and validate a demonstrator for supply disruptions or internal continuity. We start from the existing foundation and agree the capabilities, data and connections the case needs.

TRACE

Origin + consequence

ORIGIN, MINA, ambiguity control and experience with data contracts and evaluation using separated evidence.

YOUR COMPANY

Problem + technical contribution

Industrial knowledge, complementary development, authorised data and an environment for assessing the demonstrator against agreed criteria.

A joint R&D objective. Contributions, ownership of results, evaluation and conditions are defined before development.

Propose a collaboration ↗

WHAT YOU CAN TAKE AWAY

A complete guide.
A scope you can review.

Offer, engines, evidence with denominators, case requirements and collaboration process. New public material; third-party confidential documents and data are not part of these downloads.

Documentation and an explanatory model; the kit contains no installer, scientific engines or application connected to customer systems.

START WITH THE CASE

What you need to protect.
What we can assess.

Tell us about the process, disruption or decision you want to study. A general description is enough to start; we agree scope before receiving real files.

  1. 01

    Check compatibility

    A useful question, systems, formats, coverage and authorised access.

  2. 02

    Agree the work

    Data, contract, custody, deliverables, comparison, support and budget.

  3. 03

    Develop and evaluate

    A scoped case or demonstrator, results with limits and a decision on next steps.

Tell us what you need

Industrial review, adaptation or R&D collaboration.

General context only. Do not include records, credentials or other people’s personal data. This form does not accept files.

The enquiry is sent to TRACE’s support channel on Telegram. We will use your contact details to reply and follow up.

BEFORE GETTING STARTED

Practical questions.

Are these two products or two areas of Supply?

They are two applications of the same foundation. Internal operations focuses on assets, maintenance, work orders, capacity and spare parts inside the company. Logistics and supply focuses on suppliers, replenishment, routes, inbound supplies and deliveries. They overlap: a delayed spare part belongs to supply and can affect internal maintenance. We can study that connection in one case, with an agreed profile, data and evaluation.

Is Supply only about transport?

It includes internal continuity: assets, maintenance, spare parts, inventory, capacity and obligations, as well as supply and logistics dependencies. Each case needs its own profile and data.

Has it been validated in a real industrial operation?

The reviewed evidence is internal and synthetic. External compatibility preparation exists, but it does not establish an executed industrial blind cohort or realised financial usefulness.

Can I install it or upload my data here?

This page is a presentation with an illustrative model. It publishes no installer and accepts no files. A technical delivery requires agreement on licensing, security, environment and distribution scope.

Are Supply and D-IT the same?

They are separate lines. Supply has its own engines, contracts and industrial continuity studies. D-IT’s Kubernetes, OTLP and n8n demonstrations do not validate Supply.