IRX is a Java/Spring Boot dispatch optimization platform with a React control-tower dashboard for live routing, solver comparison, benchmark replay, and demo presentation.
- Spring Boot backend API for dispatch jobs, live sessions, rescue, compare, and BigData-lite ingestion.
- Hybrid optimization flow: seed race from IRX/VROOM/OR-Tools/PyVRP, IRX refinement, dominance guard, and final route trace.
- React/Vite dashboard in
playground/for live map, benchmark, demo scenario, decision trace, and API sandbox. - Packaging scripts for Windows portable release and installer-style delivery.
- Documentation for APIs, architecture, dynamic dispatch, benchmark evidence, dashboard usage, and report writing.
Install Python solver/ML runtime dependencies:
py -3-m pip install --upgrade pip
py -3-m pip install -r requirements.txt
powershell -ExecutionPolicy Bypass -File scripts\check-ml-runtime.ps1Compile backend:
.\gradlew.bat compileJava -x test --no-daemon --console=plainRun backend API locally:
.\gradlew.bat bootRun --args="--server.port=18116"Run dashboard locally:
cd playground
npm install
npm run devOpen:
http://localhost:5173
Backend compile:
.\gradlew.bat compileJava -x test --no-daemon --console=plainFrontend checks:
cd playground
npm run typecheck
npm run buildPortable baseline smoke:
.\scripts\smoke-package-baseline.ps1Full Windows portable release gate:
.\scripts\build-windows-portable-all.ps1 -RuntimeRoot C:\runtimes -SkipInstallerIf runtimes are already prepared:
.\scripts\run-windows-portable-release-gate.ps1 `-JreDir C:\runtimes\jre-21`-PythonDir C:\runtimes\python-pyvrp `-VroomExe C:\runtimes\vroom\vroom.exe`-OsrmDir C:\runtimes\osrm `-OsrmDataDir C:\runtimes\osrm-hcmcdocs/DASHBOARD_TUTORIAL.md— full dashboard usage guide.docs/REPORT_GUIDE.md— report structure, best results, screenshots, and demo evidence to include.docs/API_REFERENCE.md— API overview.docs/API_EXAMPLES.md— API examples.docs/DYNAMIC_DISPATCH.md— live/stateful dispatch concept.docs/ADAPTIVE_ML_POLICY.md— adaptive optimization policy.docs/BENCHMARKS.md— benchmark and solver comparison evidence.docs/WINDOWS_PORTABLE_RELEASE.md— Windows portable packaging and smoke plan.docs/REPO_CLEANUP.md— cleanup rules and generated-file policy.
The repo can build backend and dashboard from source. A fully portable Windows package additionally needs native runtime inputs:
- JRE 21 Windows x64.
- Portable Python with
pyvrpinstalled. - Native
vroom.exe. - Native OSRM binaries plus prebuilt
.osrmmap data.
These heavy runtime files must not be committed to git. Use docs/WINDOWS_PORTABLE_RELEASE.md for the release flow.