Skip to main content
Change Story

Algorithmic logistics optimisation

Algorithmic logistics optimisation

Simulating the minimal logistics costs based on real-life demand.

The challenge

As restrictions from the COVID-19 pandemic eased, our client's Click & Collect service faced declining demand, jeopardising its financial viability. A business case was initiated to assess the feasibility of implementing centralised fulfilment to reduce costs. Estimating real-world logistics costs proved challenging despite potential staffing efficiency and pick rate improvements. 

The approach

  • Store location analysis identified optimal clusters with short distances and high density for centralised picking trials. 
  • Driving metrics were extracted from online maps to calculate driving cost based on fuel and driver hourly rate. 
  • Existing algorithms were insufficient, so they were enhanced to account for capacity constraints and multiple pickups and deliveries. 
  • Through simulations using historical Click & Collect demand data, lowest achievable variable logistics costs were determined.

The results

Clusters were defined which identified financially viable centralised picking stores and the ideal fulfilment hub within each cluster. 

Logistics cost modelling was completed to show lowest possible logistics costs based on historical demand and the future logistics costs based on C&C demand estimates. 

The new logistics algorithm for order capacity planning, and pick-up and delivery routing was developed and integrated.

centralised picking cs footer letterbox