An optimization engine find the best combination for assigning tasks to a fleet. To compare performances we create a set up of around 600 spots, 6 task types (with different priorities) and a fleet of 8 drivers
set up for the comparison
We prepare the job file and we send it to the routific api service and visualize the solution
routific solution explorer
We can run the optimization engine from a blank system or after a routific solution.
comparison between routing engine
Starting from a routific solution we see that the optimization engine at first improves big springs
optimization improves big springs
We use the following kpi:
completion
: percentage of van capacity filledduration
: time spentpotential
: value of the vandistance
: distance of the routeThe score is calculated via:
$$ score = \frac{occupancy * potential}{duration * distance} $$
comparison of kpi between engines
Routific was sending a complete solution within 3 minutes, the optimization engine was sending a better solution in 30 minutes, to speed up the process we introduced new moves
execution time