Sener – ToD, Transportation on Demand, optimising the planning of a fleet of vehicles
Sener
Sector: Mobility
Business Case
ToD is an algorithm that optimises the planning of a vehicle fleet. It can manage the pick-up and drop-off of passengers from geographic zones based on the road network of each zone. It has been designed to handle a range of customer requests both in the future and in real time.
Objectives
The objective of ToD is to minimise fleet costs so that the vehicles’ routes are optimal while consuming the least amount of fuel possible and, at the same time, meeting the greatest number of requests possible.
Use case
ToD is a VRPTW (Vehicle Routing Problem with Time Windows) type algorithm. It plans vehicle routes taking into account the constraints imposed on the capacity problem, times and time intervals for customer visits. To do this, it is necessary to have the locations of each customer and a time matrix showing the time it takes to travel from all points to one another.
Infrastructure
On Premise (Microservicios dockerizados) y Cloud.
Technology
Automatic or Deep Learning
Data
Projects and offers developed by SENER, based on real use cases, with real customers and real needs.
Resources
Profiles in the back-end, front-end environment and applications to provide advice on the technical aspects of the different technologies used, to define user needs and features that the tool must offer, to act as a tester of the different versions as they are issued, etc. Profiles for algorithm development and software implementation.
Difficulties and learning
One of the challenges faced was effective communication and understanding between technical experts specialising in microgrids and those from different backgrounds, such as data scientists with a focus on optimisation and algorithms, as well as software developers, etc. It is key to properly dimension scalability and resource consumption. This is an important decision point due to the number of requests that the application may have to deal with and it impacts the development technically and economically.
KPIs (business impact and metrics of the model)
From a business point of view, this ToD system makes it possible to meet the same passenger demand using a smaller fleet (fewer driver-vehicles). In other words, for the same income, it allows costs to be minimised by making a more efficient use of the available resources-float. From an algorithmic point of view, for a given level of passenger-demand, there are two metrics that define performance: 1.- number of vehicle-driver hours consumed, and 2.- percentage of jobs-demand attended.
Funding
Island Council of Tenerife Own resources: SENER Mobility
