Supermarket Delivery Vehicle Routing Optimisation

The objective

Investigate methods to optimise the routes of delivery vehicles from depot to stores. This is important to Morrisons as more efficient delivery routes can lead to a reduction in both carbon emissions and transport costs.

"Morrisons' main priority is our customers. Because of our vertically integrated supply chain we have expertise and data across many components of the food systems. We understand that small, data-driven changes in e.g. the manufacturing process, all lead back to what is available on the shelf for the customer, and what price that item is available for."

The method

The Data Scientist researched the best methods for optimising delivery vehicle routes and settled on Mixed Integer Linear Programming Optimisation.

This method was applied to a mixture of real depot / store locations and anonymised product demand data. Morrisons supplied additional knowledge of the delivery system, such as the number of pallets that fit in a delivery vehicle and the limits on the length of time a driver can be driving for.

The result

Route optimisation reduced the total required travel distance by over 40%, when compared with the most basic delivery strategy.

This proof of concept work into optimisation techniques, surfaced important questions and decisions that internal colleagues need to be addressing when planning delivery vehicle routes.

The project raised new opportunities to investigate across the delivery network. The analysis into the success of optimisation fed into internal conversations and very quickly, the benefits were intrinsically combined with internal work we do to reduce costs and carbon emissions and to improve efficiency.

Knowledge exchange with a focus on methods and techniques, invaluable to daily operations and methods that are unique to Morrisons challenges as a major retailer.

Hiring

After spending six months optimising delivery routes, Jacob, the Data Scientist assigned to the project, was then hired by Morrisons upon completing the programme and continues to use data to find opportunities to reduce waste and improve availability for their customers.

Through the wider LIDA networks, Morrisons has also been introduced to other Data Scientists who have later joined their growing Data Science team, with Morrisons hiring a total of three programme alumni to date.

These team members bring value into the business via current projects such as increasing availability for customers, reducing Morrisons produced waste and explaining data clearly to the non-data literate.

University relationships & further opportunities

Morrisons have built relationships with data scientists, their supervisors and the wider LIDA team – which has led to further conversations and the opportunity to support other research projects.

During DSDP and general LIDA outreach events, members of the Morrisons Data Science team have also been introduced to other academic teams, enabling us to continue to widen our relationships within and between universities. This has led to the development of new collaborations and opportunities which continue to use data and share ideas for the wellbeing of the general public.

Working with the DSDP team was a real asset, especially knowing they had the capacity to fully explore our challenge. Balancing research and development with the immediate needs of the business can be tricky, so their dedicated focus was invaluable. The team was also well placed to collaborate with other academics (both at Leeds University and at other universities and companies across the country). This resulted in well-rounded solutions rooted in current research. Peter Laflin, Data & Analytics Director, Morrisons

Register your interest in a 6-month project