2. What ?
And Why ?
Scheduling and planning a fleets
efficient utilization.
Rising demand of delivery services.
Limiting human capabilities.
Human Error.
3. The Traveling Salesman Problem
Imagine Visiting Multiple cities.
You would want to save some fuel and visit as
soon as possible.
How would you do it ?
How did you come up with this solution ?
Were there no other solutions available ?
Can you formulate it mathematically ?
4. TSP
There are many ways to solve this
problem.
(From top to bottom)
1. Brute Force
2. Branch and Bound
3. Nearest Neighbour
5. Google OR Tools
A Multi Platform Optimization Tool
Box (C++, Python, C#, and Java).
Enables Constraint Programming.
Contains multiple Solving algorithms.
Open Source under Apache 2.0
6. Our Problem
We have a finite number of drivers with some
starting location (Depot) and some starting time.
We have a finite number of bookings with a pick
up window.
Each pick up has a drop off.
*We were also working on driver breaks constraint
but the OR Tools breaks constraint had bugs.
7. Formulating Problem
Considering each location as Node to be
visited.
All depots are out starting location.
Each Pick up (Node) location must
succeed by a Drop off (Node).
Both Pickup and Drop Off must be
visited by the same driver.
Node visiting priority is set to the
scheduled time.
Model is allowed to not visit certain
nodes if needed.