Network construction, price-gap calculations, transport scenarios, robustness analysis, corridor prioritization and normalized flow optimization were subjected to internal QA.
BOI RESEARCH · PROJECT 1
Validation
Scientific validation status of the Graph-Based Combinatorial Optimization for Agricultural Supply Networks project.
Current scientific status: analytically finalized and reproducible, but not yet empirically validated.
01 · VALIDATION FRAMEWORK
What has been validated?
The analytical workflow is documented with assumptions, scenario definitions, source provenance and reproducibility packages.
02 · VALIDATION GATES
Evidence gates
Internal consistency, deterministic calculations and quantitative quality checks were completed across the analytical workflow.
Methods, assumptions, scenarios, source registers and research packages are documented to support independent reproduction.
World Bank RTFP, INSBU price bulletins and other public evidence are integrated through explicit source and crosswalk controls.
Transaction-level freight, observed supply and demand, operating costs, logistics capacity and independent holdout observations are still required.
Transport parameters and model outputs have not been recalibrated because the required independent empirical observations are not yet sufficient.
An independent holdout dataset must be populated and evaluated before BOI can claim empirical predictive validation.
03 · EMPIRICAL REQUIREMENTS
What is still required?
- Observed transaction-level freight tariffs.
- Observed supply and demand quantities.
- Verified transport and operating cost components.
- Observed logistics and vehicle capacity.
- Verified road accessibility and condition observations.
- Independent holdout observations for out-of-sample testing.
04 · SCIENTIFIC INTEGRITY
No unsupported empirical claim
BOI does not label the model as empirically validated until the required primary observations have been collected, quality controlled, integrated and evaluated against an independent holdout dataset.
05 · CURRENT DECISION
Validation gate: HOLD
The analytical baseline is preserved. Recalibration, empirical validation and out-of-sample performance claims remain on hold until the evidence requirements are satisfied.
This distinction allows BOI to present a professional and scientifically transparent research platform without overstating the current evidence.