Full Text Available
Note: Clicking the button above will open the full text document at the original institutional repository in a new window.
An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics
Huskisson, D. 2025. An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/89adb462-000a-4d56-b29a-4dc5485cc601
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
Stellenbosch : Stellenbosch University
2025
|
| Subjects: | |
| Tags: |
No Tags, Be the first to tag this record!
|
| _version_ | 1867613949762994176 |
|---|---|
| access_status_str | Open Access |
| author | Huskisson, Dominic |
| author2 | Van Vuuren, J. H. |
| author_browse | Huskisson, Dominic Van Vuuren, J. H. |
| author_facet | Van Vuuren, J. H. Huskisson, Dominic |
| author_sort | Huskisson, Dominic |
| collection | Thesis |
| dc_rights_str_mv | Stellenbosch University |
| description | Huskisson, D. 2025. An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/89adb462-000a-4d56-b29a-4dc5485cc601 |
| format | Thesis |
| id | oai:scholar.sun.ac.za:10019.1/132214 |
| institution | Stellenbosch University (South Africa) |
| language | English |
| last_indexed | 2026-06-10T12:44:16.501Z |
| license_str | Other — see source repository |
| provenance_str_mv | Harvested via OAI-PMH from SUNScholar — Stellenbosch University Repository |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| publisher | Stellenbosch : Stellenbosch University |
| publisherStr | Stellenbosch : Stellenbosch University |
| record_format | dspace |
| source_str | SUNScholar — Stellenbosch University Repository |
| spelling | oai:scholar.sun.ac.za:10019.1/132214 An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics Huskisson, Dominic Van Vuuren, J. H. Searle, C. Van Eeden, J. Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering. Multiagent systems Delivery of goods -- Environmental aspects Carbon dioxide mitigation UCTD Huskisson, D. 2025. An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics. Unpublished masters thesis. Stellenbosch: Stellenbosch University [online]. Available: https://scholar.sun.ac.za/items/89adb462-000a-4d56-b29a-4dc5485cc601 Thesis (MEng)--Stellenbosch University, 2025. ENGLISH ABSTRACT: The availability of an effective last-mile retail delivery solution has become an urgent requirement in supply chain management due to the significant growth of the e-commerce sector and the accompanying proliferation of doorstep deliveries. Crowd logistics offers innovative solutions by involving ordinary individuals in logistics processes. In particular, crowd logistics in the context of last-mile deliveries involves employing third-party courier drivers in addition to the company-dedicated delivery feet of the retailer, tasked with order fulfilment at customers. These drivers are not associated with any specific retail depot, but are rather contracted and incentivised independently to fulfil part of the company's order deliveries. The benefit of this approach is rooted in the notion that third-party courier drivers do not have to return to a depot, thereby reducing the costs associated with the final segments of traditional delivery routes. After delivery completion, these third-party drivers may continue en route to their next destinations of choice, whether these destinations are the depots of other companies, their homes, or other locations. The aim in this thesis is to design an agent-based model capable of clustering, routing, and suggesting retail deliveries to interested third-party courier drivers, thereby unlocking the delivery potential of crowd logistics. The model integrates a traditional last-mile delivery system with the potential to utilise autonomous third-party courier drivers. The traditional last-mile delivery component consists of a company-dedicated mixed feet of vehicles serving online customers from a dedicated depot. This vehicle routing problem is cast as an integer programming problem, and solved using well-known routing heuristics. The third-party courier drivers are modelled as autonomous agents, outside the retailer's control. These drivers are each offered a set of routes with accompanying incentives, upon which they can decide whether or not to accept or reject the various routes based on self-interest. Each driver aims to maximise the incentive offered to him or her, taking into account the additional travel distance that would be incurred, and favouring deliveries that are almost along their original routes of travel. The model objective is two-fold - to minimise the total operational cost and to minimise the environmental impact of the entire delivery assignment and routing schedule (the latter measured by CO2 emissions converted to their Rand-value equivalents). The model is subjected a systematic verification process aimed at ensuring the correct functionality and integration of its components. The model is subsequently evaluated in respect of a diverse set of test scenarios in a bid to gain insight into the benefits of utilising third-party courier drivers. The metrics used to evaluate the general performance of the model are based on the quality of the solutions it returns (in terms of the aforementioned model objective), the time required to solve the model, and the percentage of customers served by third-party couriers. This evaluation is carried out in conjunction with parameter variation, a sensitivity analysis, and a scenario analysis. The model is also applied to a real-world case study which further reveals the effectiveness of crowd logistics within an existing last-mile delivery system. It is found that the inclusion of third-party courier drivers significantly reduces the total solution cost, although the success of the approach depends sensitively on the saturation level of third-party courier drivers present in the network. AFRIKAANSE OPSOMMING: Die beskikbaarheid van 'n doeltre_ende oplossing vir laaste-myl kleinhandelaewering het 'n dringende vereiste in voorsieningskettingbestuur geword as gevolg van die noemenswaardige groei van die e-handelsektor en die gepaardgaande toename in drumpel-aewerings. Skarelogistiek bied innoverende oplossings deur gewone individue by logistieke prosesse te betrek. In die besonder behels skarelogistiek in die konteks van laaste-myl aewerings die indiensname van derdeparty-koerierbestuurders bykomend tot die maatskappy-toegewyde aeweringsvloot van die kleinhandelaar met die opdrag om bestellings by kliente af te lewer. Hierdie bestuurders word nie met enige spesi_eke kleinhandeldepot geassosieer nie, maar word eerder onafhanklik gekontrakteer en aangespoor om 'n deel van die maatskappy se aewrings by kliente waar te neem. Die voordeel van hierdie benadering l^e in die feit dat derdepartykoerierbestuurders nie na 'n depot hoef terug te keer nie, en sodoende die koste verbonde aan die _nale segmente van tradisionele aeweringsroetes spaar. Nadat aewering voltooi is, kan hierdie derdepartybestuurders verder na volgende bestemmings van hul keuse reis, of hierdie bestemmings die depots van ander maatskappye, hul huise, of ander liggings is. Die doel in hierdie tesis is om 'n agent-gebaseerde model te ontwerp wat in staat is om kleinhandela ewerings te groepeer, te roeteer en aan belangstellende derdeparty-koerierbestuurders voor te stel, om sodoende die aeweringspotensiaal van skarelogistiek te ontsluit. Die model integreer 'n tradisionele laaste-myl aeweringstelsel met die potensiaal om outonome derdepartykoerierbestuurders te gebruik. Die tradisionele laaste-myl aeweringskomponent bestaan uit 'n maatskappy-toegewyde gemengde vloot voertuie wat aanlyn-kliente vanaf 'n toegewyde depot bedien. Hierdie voertuigroeteringsprobleem word as 'n heeltallige programmeringsprobleem geformuleer en met behulp van bekende roeteringsheuristieke opgelos. Die derdepartykoerierbestuurders word as outonome agente buite die beheer van die kleinhandelaar gemodel leer. Hierdie bestuurders word elk 'n versameling roetes met gepaardgaande aansporings aangebied, waarop hulle, gebaseer op eiebelang, kan besluit of hulle die verskillende roetes aanvaar of nie. Elke bestuurder poog om die aansporing wat aan hom of haar aangebied word, te maksimeer met inagneming van die bykomende reisafstand wat aangegaan word, en het 'n voorkeur vir aewerings wat amper langs hul oorspronklike reisroetes gelee is. Die doel van die model is tweeledig | om die totale bedryfskoste te minimeer en om die omgewingsimpak van die algehele aeweringsopdrag en roeteskedule te minimeer (laasgenoemde gemeet in terme van CO2-vrylating omgeskakel na 'n Randwaarde-ekwivalent). Die model word aan 'n sistematiese veri_kasieproses onderwerp wat daarop gemik is om die korrekte funksionaliteit en integrasie van die komponente daarvan te verseker. Die model word vervolgens in die konteks van 'n diverse versameling toets-scenario's geevalueer in 'n poging om insig oor die voordeel van die gebruik van derdeparty-koerierbestuurders te bekom. Die maatstawwe wat gebruik word om die algemene prestasie van die model te evalueer, is gebaseer op die kwaliteit van die oplossings wat dit lewer (in terme van die bogenoemde modeldoel), die tyd wat nodig is om die model op te los, en die persentasie kliente wat deur derdepartykoeriers bedien word. Hierdie evaluering vind plaas in tandem met parametervariasie, 'n sensitiwiteitsanalise en 'n scenario-analise. Die model word ook op 'n werklike gevallestudie toegepas wat die doeltre_endheid van skarelogistiek binne 'n bestaande laaste-myl aeweringstelsel verder openbaar. Daar word bevind dat die insluiting van derdeparty-koerierbestuurders die totale oplossingskoste aansienlik verminder, hoewel die sukses van die benadering sensitief afhang van die versadigingsvlak van derdepartykoerierbestuurders teenwoordig in die netwerk. Masters 2025-05-30T06:56:16Z 2025-05-30T06:56:16Z 2025-03 Thesis https://scholar.sun.ac.za/handle/10019.1/132214 en Stellenbosch University xxiv, 190 pages : illustrations application/pdf Stellenbosch : Stellenbosch University |
| spellingShingle | Multiagent systems Delivery of goods -- Environmental aspects Carbon dioxide mitigation UCTD Huskisson, Dominic An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics |
| title | An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics |
| title_full | An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics |
| title_fullStr | An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics |
| title_full_unstemmed | An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics |
| title_short | An agent-based model for fuel and CO2 emission costs in multi-modal last-mile logistics |
| title_sort | agent based model for fuel and co2 emission costs in multi modal last mile logistics |
| topic | Multiagent systems Delivery of goods -- Environmental aspects Carbon dioxide mitigation UCTD |
| url | https://scholar.sun.ac.za/handle/10019.1/132214 |
| work_keys_str_mv | AT huskissondominic anagentbasedmodelforfuelandco2emissioncostsinmultimodallastmilelogistics AT huskissondominic agentbasedmodelforfuelandco2emissioncostsinmultimodallastmilelogistics |