Export to KML — Export2KMLExample

This example demonstrates a practical “last mile” step after solving a tour optimization: exporting the resulting optimization container to a KML file so you can visualize routes and locations in map tools such as Google Earth or other KML-capable viewers.

Unlike purely console-based examples, this one produces a tangible artifact on disk: myopti.kml.

Links

• Example source (GitHub):
  Export2KMLExample.java

What you get when you run it

1. The optimizer runs asynchronously.
2. Progress/status/error callbacks write messages to stdout.
3. When the run finishes, onAsynchronousOptimizationResult(...) is invoked.
4. The example exports the result container to myopti.kml via EntityKMLExporter.

In other words: Solve → Receive result → Export → Open in a map tool.

Execution flow (read this like a story)

1) main(...) launches the scenario

main(...) simply instantiates the example and calls example().

2) example() wires the run end-to-end

The orchestration is intentionally compact:

• Apply a license via ExampleLicenseHelper.setLicense(this);
• Register solver properties via setProperties()
• Register problem elements via addNodes() and addResources()
• Start the run via startRunAsync().get() (the blocking get() is there to keep the JVM alive until completion)

Note: The class also overrides onAsynchronousOptimizationResult(...), which is where the export happens. The get() call ensures the asynchronous run has time to finish and invoke that callback.

What is being optimized

Nodes (time-windowed geo points)

The example creates a shared list of opening hours for two days in March 2020 (Europe/Berlin time zone):

• March 6: 08:00–17:00
• March 7: 08:00–17:00

All nodes share:

• visitDuration = 20 minutes
• importance is set to 1

Nodes added (as TimeWindowGeoNode):

• Koeln
• Koeln1 (same coordinates as Koeln; useful for testing repeated/duplicate locations)
• Oberhausen
• Essen
• Heilbronn
• Stuttgart
• Wuppertal
• Aachen

Resources (capacity resources)

Two resources are created as CapacityResource, both starting at Aachen coordinates:

• Jack
  • max working time: 8 hours
• John
  • max working time: 14 hours

Shared constraints:

• max distance per working day: 1200 km
• working hours are defined by getDefaultWorkingHours() (two days; the second day extends to 20:00)

Each resource is assigned a cost matrix entry via setCost(0, 1, 1).

Solver properties (what the run is “allowed” to do)

setProperties() registers a small set of properties that define solver effort and runtime behavior:

• JOptExitCondition.JOptGenerationCount = 2000
• JOpt.Algorithm.PreOptimization.SA.NumIterations = 100000
• JOpt.Algorithm.PreOptimization.SA.NumRepetions = 1
• JOpt.NumCPUCores = 4
• JOptLicense.CheckAutoLicensce = FALSE

These are typical “developer friendly” defaults: enough compute to produce a meaningful result, with bounded runtime behavior in common environments.

Observability: progress, status, warnings, errors

This example shows two styles of progress reporting:

• onProgress(String winnerProgressString) prints a progress string
• onProgress(IOptimizationProgress rapoptProgress) is implemented but intentionally left empty

Additionally:

• onStatus(...) prints status code + message
• onError(...) prints error code + message
• onWarning(...) is present (stubbed)

Practical takeaway: you can scale observability from “just print progress” to “feed structured telemetry” without changing your modeling logic.

The KML export (the point of this example)

Where the export happens

The file is created in:

• onAsynchronousOptimizationResult(IOptimizationResult rapoptResult)

Core logic:

• Print the result (System.out.println(rapoptResult);)
• Create the output filename: myopti.kml
• Export via:
  • IEntityExporter exporter = new EntityKMLExporter();
  • exporter.export(rapoptResult.getContainer(), new FileOutputStream(kmlFile));

What is exported

The exporter writes the optimization container from the result:

• rapoptResult.getContainer()

This is important: the export uses the container that holds the solved state, not just a textual summary.

How to use the output

Once generated, open myopti.kml in a KML viewer (for example, Google Earth) to visually inspect:

• node locations,
• computed tours/routes (depending on exporter capabilities and result structure).

Implementation notes (recommended improvements)

This example is intentionally minimal. In production or reusable tooling, consider:

1. Use try-with-resources for the output stream:
   • ensures the stream is closed even on exceptions.
2. Make the output path explicit (e.g., configurable output directory) rather than relying on the process working directory.
3. Export on demand:
   • e.g., export only when a run is “good enough”, or export multiple snapshots (best-of) during progress callbacks.

Small curiosity

The class-level string returned by toString() says it saves to a JSON file, while the implementation exports to KML. Treat the Java source as authoritative; the string is likely a leftover from a similar “export” example variant.

How to run

Run the main(String[] args) method of Export2KMLExample.

After completion, check your process working directory for:

• myopti.kml