Gediminas Morkevičius aka @l3pp4rd
At DATA-DOG
In partnership with Satalia
We solve hard problems
We are developing sofware for grocery supply chain
Which compared to Maxima in baltic states
is at least 5 times bigger
has around 200 depots in the region we work with
What are we doing there?
and these are hard problems
Vehicle routing problems
How many possible routes there may be within 24 points?
Other routing problems
First of all, you have to lower the number of possibilities
and make trade offs
A similar solution to our - Oracle ORS uses a Super-PC with 400 CPUs
What do we schedule?
Grocery and other deliveries
Warehouse distribution
Imagine a schedule..
Lets make it harder
Every drop may be allocated to a time slot
main scheduling operations
Currently we have around several methods of optimization
customized ILP algorithm, but everywhere there are trade offs
cannot lock actively updated schedule for long running optimization
probably every depot should have a different kind of optimization
Maybe Ant colony optimization algorithms
there are many interesting things we will try
Hopefully producing similar results
Slides are available at: slides.gediminasm.org
Powered by: Revealjs