Why one waste collection model does not fit every area

Should recyclable waste be collected directly by truck, taken to shared drop-off points, or managed through a combination of both? Research by MSc student Alting von Geusau developed a model for KOLEKT’s Waste Simulator that compares these options area by area, showing how the lower-cost choice can change with local conditions.

Why does the way waste is collected affect system costs?

Waste collection depends on geography. Population density, the amount of waste generated, the distance to sorting facilities and the collection technology all affect how many vehicles, collection points and working hours a system requires.

This means that a collection system that works well in one place may be much more expensive somewhere else. A densely populated urban area, for example, creates a different collection challenge from a sparsely populated rural one.

Alting von Geusau’s research therefore asked how high-resolution spatial data could be used to compare different ways of collecting waste within KOLEKT’s Waste Simulator. The aim was to give planners a stronger basis for deciding which collection approach makes sense in different locations.

The key question was not which collection system is best overall, but where each system makes sense.

Why compare trucks and drop-off points area by area?

The research compares three collection systems.

In a truck-only system, vehicles collect waste directly from households. In a drop-off-point system, people bring eligible waste to shared collection locations, from where it is collected and transported onwards. This can reduce the amount of local truck activity, but it also requires investment in containers and collection-point infrastructure.

The third option is a hybrid system. Instead of applying one method across an entire country, the model compares the cost of truck collection and drop-off points within each collection area and selects the lower-cost option.

The framework was tested across six countries, covering more than 355,000 spatial cells and 859 collection areas.

This makes it possible to treat collection as a local planning decision rather than a single national choice.

What did the six-country analysis show?

In the main six-country scenario, applying truck collection across all 859 areas produced an estimated annual cost of about $860 million. Using drop-off points everywhere was much more expensive, at about $3.92 billion a year, mainly because of the infrastructure required for the drop-off network.

A hybrid approach — choosing the lower-cost option for each area — reduced the estimated cost slightly to about $850 million a year. The model selected drop-off points in 70 areas and truck collection in the remaining areas.

The difference between truck-only and hybrid collection was therefore modest across the six-country model as a whole. But the location of those differences matters: the analysis identified specific areas where drop-off points could be more economical instead of assuming that the same system should be used everywhere.

For planners, the finding suggests that the more useful question is not “Which collection method should a country use?” but “Where could a different collection method make sense?”

What did the research show about sorting-facility locations?

Alting von Geusau also examined the locations of Material Recovery Facilities (MRFs) — facilities where collected waste is sorted before it moves on for further processing.

The research compared the assigned MRF locations with a mathematical benchmark that minimises the distance waste needs to travel. Across the model, the existing locations were already relatively close to that benchmark. Moving them to the mathematical optimum improved the waste-weighted transport-distance measure by about 2%, while the corresponding effect on hybrid collection costs was below 1%.

This suggests that redesigning the location of every sorting facility would not necessarily produce large savings. In some areas, closer examination or additional facilities may be worthwhile, but the analysis points towards targeted changes rather than wholesale relocation.

The result helps planners distinguish between parts of the system where redesign may matter and those where existing assumptions are already reasonably strong.

What does this add to waste-system decision-making?

For KOLEKT, the research suggests that collection technology should be modelled at collection-area level rather than imposed as one national assumption.

This creates a more useful decision layer in the Waste Simulator. Planners can compare factors such as annual cost, cost per tonne, fleet requirements and drop-off infrastructure across different collection approaches, while areas where the difference is small can be flagged