- Researcher: Fabian Ying
- Academic Supervisors: Sam Howison, Mariano Beguerisse and Mason Porter (UCLA)
- Industrial Supervisors: Alisdair Wallis, Trevor Sidery and Robert Armstrong
The aim of this project is to develop a model that can be used to describe, predict and reduce congestion in UK supermarkets.
The grocery retail market is a highly competitive sector and thus anything that improves the store for customers is of great interest to retailers such as Tesco. Identifying and reducing congestion will both enhance the shopping experience for customers and reduce the cost to fulfill online orders from within the store.
We use anonymised mobility data inside supermarkets to develop and calibrate our models.
To predict congestion under new store layouts, we first need a model that can predict how customers move in new store layouts. To do this, we use human mobility models, which have been used to describe and predict traffic flow in cities, commuting flow between states/counties, and trade flows between countries. With these models, we predict the mobility flow of customers between any two zones in the store based on the popularity of items in the zones and the distance between the zones.
We infer the empirical mobility flow from anonymised mobility data, and we find that two of the models (the gravity and extended radiation model) can successfully predict 65% to 70% of the empirical mobility flow.
From the mobility flow, we can then estimate the number of visits (assuming shortest walk routing between purchases) that a node receives and calculate measures of congestion based on this. We also find that the two models give excellent agreement with the number of visits that are estimated from the empirical flow (see figure below).
Finally, we use a simple optimisation algorithm (simulated annealing) to find store layouts with less congestion. We assume that the basic geometry of the store is fixed and that we can only swap zones. We find that our optimised store layouts have up to 50% lower congestion values and disperse popular zones across the store (see figure below).
Hence our work has shown promising results for using application of human mobility models to predict congestion and using simple optimisation algorithm to find store layouts with less congestion.
In future, we would like to validate our results for more stores, as our prediction results have only been validated on two stores so far.
Furthermore, to estimate the number of visits, we assumed that customers walk shortest path between zones, which is likely unrealistic. In future work, we aim to build a more realistic routing model for how customers move between purchases.
Finally, we aim to incorporate business constraints into our optimisation of store layout, so we can suggest viable store layouts.