An Android navigation framework treats the shortest path method as a plug in, so the rest of the app never has to care which one runs.
In most navigation apps, the route calculation is welded into everything else. The map screen calls it, the data layer feeds it, and the drawing code reads its output directly. Trying a different algorithm then means touching half the project.
A framework designed for the Google Android platform takes the opposite approach. It treats the pathfinder as one replaceable piece behind a common interface, and uses the Abstract Factory pattern so the application can choose an implementation at runtime instead of at compile time.
The team built a working prototype in Java and ran it on a Google Nexus 6 with Android 7.0, using a campus road network of 134 nodes and 154 edges. Both algorithms returned the same routes, as expected. The interesting difference was in time.
That works out to more than 60 percent faster on average for a single route, with the largest gains in the 71 percent range. The heuristic is not free though. Estimating distances with the Euclidean formula added about 0.26 ms of overhead per query, and the authors note that Haversine would give a more accurate estimate over longer distances.
Pinandito, A., Kharisma, A.P., Perdana, R.S. (2018). Framework Design for Map-Based Navigation in Google Android Platform. Journal of Telecommunication, Electronic and Computer Engineering (JTEC), 10(1-8).
