The Veeroute optimization engine rapidly solves scheduling and planning problems such as those that occur in last mile delivery, fleet management, long haul delivery and managing field service engineers.
The optimization engine’s main component is a mathematical core which solves combinatorial problems. It finds an optimal combination which satisfies the constraints and requirements in a finite discrete search space.
The mathematical core includes the following components:
This module offers more than 100 algorithms which solve combinatorial problems. It includes well-known algorithms such as simulated annealing, ruin-and-recreate and 2-opt as well as proprietary algorithms that solve specific problems. It also includes proprietary clustering algorithms which consider details such as roads, traffic and obstacles.
These models offer different route calculation options. Some account for the
penalties you’ll incur if you don’t meet customer requirements such as delivery
time windows. Others account for time-based variations such as the time your
drivers spend in transit or at a warehouse loading or unloading vehicles and
Configurations are combinations of algorithms and routing models. They also
include information about how to apply algorithms and routing models to solve
a problem. The optimization engine automatically selects the best configuration
for your problem and your dataset.
A configuration could include rules such as:
- Run a set of selected algorithms several times.
Since there is randomness in the algorithms,
the same algorithm may produce different results on the same input dataset.
- Choose the 5 best results.
- Use the best result as an input dataset for another set of
algorithms and try to improve it within 10 minutes.