Optimisation of Residential Waste Collection
Balancing Travel Time and Visual Attractiveness in Side Loader Routes
A.E.N. van der Helm (TU Delft - Civil Engineering & Geosciences)
B. Atasoy – Graduation committee member (TU Delft - Transport Engineering and Logistics)
A.J. Pel – Graduation committee member (TU Delft - Transport, Mobility and Logistics)
Kristian Hauge – Mentor (AMCS)
Koen van Duurling – Mentor (AMCS)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Residential waste collection with side loaders requires double traversal because vehicles service only one kerbside per pass. Consequently, routes optimised for efficiency often result in fragmented service of areas, long time intervals between visiting opposite kerbsides, and complex route shapes that are unintuitive to execute. Therefore, study developed a single-route optimisation framework that balances travel time efficiency with visual attractiveness under side loader operating constraints, including traffic directions, double traversal, and U-turn restrictions. The problem was formulated as an asymmetric travelling salesman problem (ATSP) on a directed network, where nodes represent directed kerbside street segments and travel times are computed as the fastest paths on the underlying road network. Visual attractiveness was quantified using complementary metrics for regional compactness, local compactness, and route complexity. These metrics were defined on an abstract network using straight-line connections to maintain computational tractability. They were combined with travel time in a weighted sum objective and optimised using simulated annealing. This enables a configurable trade-off between visual attractiveness and efficiency. The framework was evaluated on 12 municipal case studies provided by AMCS. Compared to the travel-time-optimal routes generated by AMCS, the framework consistently reduced neighbourhood fragmentation, revisit spans between opposite kerbsides, and intra-route crossings, while moderately increasing total travel time and directed Euclidean distance. A sensitivity analysis showed that uniformly scaling the visual attractiveness weights controls the trade-off between visual attractiveness and efficiency. However, very small scaling factors produced less consistent improvements in visual attractiveness, whereas larger factors yielded diminishing improvements while travel time increased substantially. Overall, the results suggest that visual attractiveness can be formalised on an abstract network and integrated into a configurable optimisation framework to produce visually attractive side loader routes on the road network.