Inspiration
We are solving the problem of fleet optimization for any Hyperlocal business. The pain that we tried to address is, once a delivery executive has made the final delivery in his current batch, where should he be routed next. In an ideal or happy scenario, the executive should be routed to the nearest point of presence of the business that also has some demand. But in a “non-ideal” world, this heuristic does not work well as the nearest neighbourhood may not have an adequate demand to be provided with the supply of delivery executives.
What it does
We have modeled this as an optimization problem where the “average wait” time for all orders is the variable to be minimized while having constraints on the “average distance” that a delivery executive has to travel to reach the recommended destination.
How I built it
The approach is there in the google doc shared herewith
Challenges I ran into
Due to the anonymization of the data, some of the more relevant relationships lost their statistical significance. This led us to spend a lot of time in mocking the scenarios using the partial data we had
Log in or sign up for Devpost to join the conversation.