Freight demand
A representation of transport demand should account for the movement of goods, not only people. That includes freight transport and last-mile deliveries.
This matters for the IMP because freight and passenger vehicles share the same roads. A flooded link delays both, and the economic consequences of blocked freight are often larger than those of delayed commuters.
How freight demand is estimated
Freight transport
Freight demand can be estimated from the spatial distribution of:
- Economic activities
- Logistics facilities
These provide information on the quantity of goods movements and their origins and destinations.
Last-mile delivery
At the urban scale, last-mile delivery demand is increasingly associated with retail and e-commerce activity. It can be estimated using:
| Input | Contribution |
|---|---|
| Population and household characteristics | Who is receiving deliveries |
| Land use | Where they are |
| Purchasing behaviour | How often they order |
| Observed or assumed parcel-delivery rates | How that converts into drops |
From estimates to vehicle movements
These estimates are translated into shipments, and subsequently into vehicle tours, by assigning goods to:
- Distribution centres or depots
- Delivery locations
- Vehicle types
- Departure times
- Routes
In an agent-based framework, freight carriers, delivery vehicles or shipments can therefore be represented alongside individual travellers, allowing passenger and goods movements to interact on the same transport network.
Freight in MATSim
In MATSim, freight and last-mile delivery demand can be represented within the same activity-based demand framework used for passengers.
Where freight activities are generated by an external demand or logistics model, such as MILES or MATRA, those activities can be translated into MATSim-compatible activity plans.
A delivery van is just another agent
Each freight vehicle is treated as an individual agent with a daily plan consisting of a sequence of activities and connecting trips:
| Plan element | What it represents |
|---|---|
| Activity | Departure from a depot |
| Activity | Deliveries or collections at different locations |
| Activity | Return to the depot |
| Connecting legs | The corresponding vehicle movements through the transport network |
This approach enables externally estimated freight and last-mile delivery demand to be incorporated into the same activity-plan structure as passenger demand, and simulated together.
MILES
MILES (Multimodal Integrated Logistic Evaluation Software) is a Java-based package of algorithms and data structures specialised for modelling logistics.
Its algorithms cover:
- Short-haul freight
- Long-haul freight
- Dynamic routing
- Mixed fleet optimisation
Inputs and outputs to be supplied
The handbook lists Inputs and Outputs for MILES as headings with no content
beneath them. Neither has been specified yet.
MATRA
MATRA is a Python-based computational modelling framework developed to investigate the operation and transition of road-freight logistics systems under increasing transport electrification.
How it works
The model combines three methods:
| Method | Role |
|---|---|
| Agent-based modelling | Represents the actors and their decisions |
| Network-based routing | Moves vehicles over real road geometry |
| Mathematical optimisation | Finds efficient fleet and charging arrangements |
It represents interactions between:
- Freight depots
- Vehicles
- Customers
- Public charging infrastructure
- Electricity substations
How it is parameterised
MATRA is parameterised using empirical logistics and operational data, which characterise:
- Depot activity
- Customer demand
- Delivery patterns
- Fleet operations
Synthetic and hypothetical elements are introduced where empirical information is unavailable.
What it evaluates
This enables MATRA to evaluate how alternative arrangements affect system performance:
| Lever varied | Outcome assessed |
|---|---|
| Fleet compositions | Logistics performance |
| Charging strategies | Energy demand |
| Infrastructure configurations | Operational feasibility |
| Electrification scenarios | Infrastructure utilisation |
Its role in the IMP
Although MATRA is fundamentally a freight transport and logistics model, within the scope of this project it is used as a transport-demand tool, providing freight-related activity and demand information to the wider modelling framework.
MATRA is documented twice
Further details on MATRA, including its modelling approach, required inputs and generated outputs, are given in the transport model chapter. This page covers only its role as a source of freight demand.