Data-driven analysis of urban logistics policies based on vehicle traces in the Copenhagen capital region

Sebastian Hörl, Aalae Benki

10 June 2025

hEART 2025

• Cities are increasingly trying to set up urban logistics policies
• Sustainable Urban Logistics Plans, also the case for Copenhagen
   
• Goal is to set up a comprehensive framework for simulating city logistics
   
• Copenhagen has an impressive network of logistics operators
• Which allows us to base our model assumptions on real data
   
• This is a first step in this process, presenting a data-driven model

Introduction

• Stop sequences by coordinate (caveat: no depots)
• 9 operators (5 food & beverage, 4 parcels & postal)
• Each for about 2 weeks, in the Copenhagen area

Available data

• Stop sequences by coordinate (caveat: no depots)
• 9 operators (5 food & beverage, 4 parcels & postal)
• Each for about 2 weeks, in the Copenhagen area

Available data

• Data-driven model to assess impact of policies
• Derive information form direct data processing
   
• Policy 1: Enforcement of electric vehicles
• Policy 2: Implementing a logistics hub

Data-driven model

First step: routing

• All stop sequences have been routed on the road network
   
• Allows deriving (lower bound) driven distance
   
• Allows deriving (approximately) emissions
   
• Show daily flow on a map

Baseline

• Analysis on three areas:
  • Environmental zone of Copenhagen
  • Inner city of Copenhagen
  • Medieval city center

• Policy: Implement a driving restriction zone for non-electric vehicles
   
• Perimeter: Either of the three analysis zones
   
• Concept:
  • Tag all vehicles that have at least one stop inside the zone as electric
  • All routes of the selected vehicles become electric
     
• Caveat:
  • In reality, operators would probably reassign their vehicle fleet and reduce the number of required electric vehicles

Electric vehicle zone

• Enforcing an electric vehicle zone has network effects beyond the selected zone
   
• Implementing the Medieval EVZ

Electric vehicle zone

• Map of zone 1, 2, 3

• Enforcing an electric vehicle zone has network effects beyond the selected zone
   
• Implementing the Inner EVZ

Electric vehicle zone

• Map of zone 1, 2, 3

• Enforcing an electric vehicle zone has network effects beyond the selected zone
   
• Implementing the Inner EVZ

Electric vehicle zone

• Map of zone 1, 2, 3

Logistics hub

• Policy: Implement micro-hub through which all goods with a specific destination area need to pass. The final delivery is done using zero-emission vehicles.
   
• Concept:
  • Identify all stops of all vehicles in the delivery zone.
  • Replace the first visit along each tour of such a stop with the hub location.
  • Ignore all other stops in the area.
     
• Caveats:
  • Tour information in our data is limited.
  • We don't calculate last mile operations (more complex problem).

• Map of the hub and operating area

Logistics hub

• Policy: Implement micro-hub through which all goods with a specific destination area need to pass. The final delivery is done using zero-emission vehicles.
   
• Concept:
  • Identify all stops of all vehicles in the delivery zone.
  • Replace the first visit along each tour of such a stop with the hub location.
  • Ignore all other stops in the area.
     
• Caveats:
  • Tour information in our data is limited.
  • We don't calculate last mile operations (more complex problem).

• Provide some results
   
• Show difference map of flows

Overview

• Provide some results
   
• Show difference map of flows

Overview

• Provide some results
   
• Show difference map of flows

Overview

• Provide some results
   
• Show difference map of flows

Overview

• We will loose all the data :)
• Set up an overall processing pipeline

Next steps

Next steps

Next steps

Next steps

Next steps

Questions?

https://slides.com/sebastianhorl/heart25