O objetivo deste trabalho é apresentar um método que possibilite a avaliação e a comparação de áreas urbanas quanto à dificuldade imposta pela malha viária à circulação dos veículos que realizam a distribuição de última milha Busca-se um método automatizado que seja global, ou seja, que possa ser aplicado em qualquer cidade do mundo, não dependa de dados de campo nem de informações cujo acesso é restrito. A relevância deste trabalho é justificada pela carência de ferramentas que auxiliem a tomada de decisões relacionadas a políticas de transporte e planejamento da distribuição de última milha em cidades. Primeiro, é feita uma revisão bibliográfica de métodos e métricas de avaliação de cidades e redes viárias e de transportes, a fim de identificar quais podem ser adaptadas para o problema da distribuição de carga urbana. Em seguida é apresentado o procedimento do método proposto, que contempla as duas partes da distribuição de última milha: circulação local e acesso/regresso ao distrito de entrega. O método avalia a malha viária quanto a influência da sua infraestrutura nas distâncias das viagens através de um índice de \"retilinidade\" baseado em uma métrica revisada da literatura, e avalia a influência da ocupação viária nos tempos de viagem através das velocidades médias estimadas em dados históricos. O método ainda contempla a possibilidade do uso de veículos não motorizados nas entregas, adicionando a topografia da área urbana estudada à avaliação realizada. Os dados de entrada necessários para a aplicação do método são estabelecidos pelo usuário com base na região e no tipo de cadeia de suprimentos que se deseja estudar e as informações necessárias para a análise são obtidas através das ferramentas Google Maps Directions API e Open Street Map, cumprindo os objetivos de generalidade e automação desejados. Para a parte da circulação local, são apresentados resultados da aplicação do método proposto em seis áreas urbanas de seis cidades diferentes (São Paulo, Rio de janeiro, Bogotá, Nova Iorque, Londres e São Francisco), e, para a parte do acesso/regresso ao distrito de entrega, são apresentados resultados para quatro potenciais locais de Centro de Distribuição e uma área de entrega em São Paulo. Os resultados consistem em uma coletânea de métricas, gráficos e mapas que permitem a avaliação e comparação das áreas urbanas estudadas, como proposto. Adicionalmente são apresentadas as correlações entre as métricas estruturais das malhas viárias e as métricas associadas ao desempenho da distribuição de última milha, a fim de identificar possíveis causalidades entre elas. Por fim, os resultados obtidos são discutidos e são sugeridos próximos passos para a continuação deste trabalho. / The goal of this thesis is to introduce a method that allows the evaluation and comparison of urban areas regarding the difficulty imposed by the road network to the circulation of vehicles that perform the last mile distribution. The method must be automated and global, meaning that it can be applied in any city and does not rely on field data or restricted information. This thesis is motivated by the scarcity of supporting tools for elaborating transport policies and planning the last mile distribution operations in cities. Firstly, a literature review is conducted to identify urban and road network metrics that can be applicable to urban freight distribution analysis. Next, the proposed method is presented. It analyses two parts of the last mile distribution: local distribution and line-haul. The method evaluates the influence of the road network infrastructure on travel distances through a straightness index based on metric found in the literature review, and it evaluates the influence of road occupation on travel times through estimated speeds based on historical data. All inputs required for the application of the method are defined by the user and are based on the characteristics of the study area and type of supply chain. The information required for the analysis is obtained through the Google Maps Directions API and Open Street Map tools, which fulfill the automation premise of the method. For the local distribution, a study case is conducted in six urban areas from six cities (Sao Paulo, Rio de Janeiro, Bogota, New York, London and San Francisco). For the line-haul, the case study includes four potential distribution center locations and a delivery area in Sao Paulo. The results are a collection of metrics, graphs and maps that allow the evaluation and the comparison of the six urban areas, as proposed. Additionally, the correlations between the structural metrics of the road networks and the drivers of the last mile distribution are presented, in order to identify possible causalities between them. Finally, the results are discussed and further steps for this work are suggested.

O objetivo deste trabalho é apresentar um método que possibilite a avaliação e a comparação de áreas urbanas quanto à dificuldade imposta pela malha viária à circulação dos veículos que realizam a distribuição de última milha Busca-se um método automatizado que seja global, ou seja, que possa ser aplicado em qualquer cidade do mundo, não dependa de dados de campo nem de informações cujo acesso é restrito. A relevância deste trabalho é justificada pela carência de ferramentas que auxiliem a tomada de decisões relacionadas a políticas de transporte e planejamento da distribuição de última milha em cidades. Primeiro, é feita uma revisão bibliográfica de métodos e métricas de avaliação de cidades e redes viárias e de transportes, a fim de identificar quais podem ser adaptadas para o problema da distribuição de carga urbana. Em seguida é apresentado o procedimento do método proposto, que contempla as duas partes da distribuição de última milha: circulação local e acesso/regresso ao distrito de entrega. O método avalia a malha viária quanto a influência da sua infraestrutura nas distâncias das viagens através de um índice de \"retilinidade\" baseado em uma métrica revisada da literatura, e avalia a influência da ocupação viária nos tempos de viagem através das velocidades médias estimadas em dados históricos. O método ainda contempla a possibilidade do uso de veículos não motorizados nas entregas, adicionando a topografia da área urbana estudada à avaliação realizada. Os dados de entrada necessários para a aplicação do método são estabelecidos pelo usuário com base na região e no tipo de cadeia de suprimentos que se deseja estudar e as informações necessárias para a análise são obtidas através das ferramentas Google Maps Directions API e Open Street Map, cumprindo os objetivos de generalidade e automação desejados. Para a parte da circulação local, são apresentados resultados da aplicação do método proposto em seis áreas urbanas de seis cidades diferentes (São Paulo, Rio de janeiro, Bogotá, Nova Iorque, Londres e São Francisco), e, para a parte do acesso/regresso ao distrito de entrega, são apresentados resultados para quatro potenciais locais de Centro de Distribuição e uma área de entrega em São Paulo. Os resultados consistem em uma coletânea de métricas, gráficos e mapas que permitem a avaliação e comparação das áreas urbanas estudadas, como proposto. Adicionalmente são apresentadas as correlações entre as métricas estruturais das malhas viárias e as métricas associadas ao desempenho da distribuição de última milha, a fim de identificar possíveis causalidades entre elas. Por fim, os resultados obtidos são discutidos e são sugeridos próximos passos para a continuação deste trabalho. / The goal of this thesis is to introduce a method that allows the evaluation and comparison of urban areas regarding the difficulty imposed by the road network to the circulation of vehicles that perform the last mile distribution. The method must be automated and global, meaning that it can be applied in any city and does not rely on field data or restricted information. This thesis is motivated by the scarcity of supporting tools for elaborating transport policies and planning the last mile distribution operations in cities. Firstly, a literature review is conducted to identify urban and road network metrics that can be applicable to urban freight distribution analysis. Next, the proposed method is presented. It analyses two parts of the last mile distribution: local distribution and line-haul. The method evaluates the influence of the road network infrastructure on travel distances through a straightness index based on metric found in the literature review, and it evaluates the influence of road occupation on travel times through estimated speeds based on historical data. All inputs required for the application of the method are defined by the user and are based on the characteristics of the study area and type of supply chain. The information required for the analysis is obtained through the Google Maps Directions API and Open Street Map tools, which fulfill the automation premise of the method. For the local distribution, a study case is conducted in six urban areas from six cities (Sao Paulo, Rio de Janeiro, Bogota, New York, London and San Francisco). For the line-haul, the case study includes four potential distribution center locations and a delivery area in Sao Paulo. The results are a collection of metrics, graphs and maps that allow the evaluation and the comparison of the six urban areas, as proposed. Additionally, the correlations between the structural metrics of the road networks and the drivers of the last mile distribution are presented, in order to identify possible causalities between them. Finally, the results are discussed and further steps for this work are suggested.

Region Östergötland are responsible for deliveries of consumable goods to hospitals, health centres, public dental clinics and other customers in Östergötland county. Today, this is done through two different material flows. These are the inner and the outer material flows. Deliveries from the inner material flow pass through a central warehouse located in Norrköping, before they are delivered to customers, whereas deliveries in the outer material flow are delivered straight from an external supplier to the customer. Today, Region Östergötland wants to gain a better control over the material flow of consumable goods. They also want to investigate how a more coordinated freight distribution, and therefore a greater inner material flow, would affect their own business, seen from a holistic perspective.  The purpose of this study was to investigate how the handling of goods, the delivery service and the total cost of ownership would be affected by a more coordinated freight distribution of consumable goods to the health care system within Region Östergötland. This to provide a basis for future decision making regarding the material flow of consumable goods. The conclusions drawn are mainly based on the empirical studies that were carried out during the study. The empirical studies were collected in various ways and were of qualitative, as well as quantitative, nature. Interviews were conducted with employees from Region Östergötland as well as with employees from some contracted companies. Delivery protocols, as well as surveys, were handed out to and answered by customers in the health care system. Data from Region Östergötland’s enterprise resource planning systems, as well as contracts, were also examined. The reports frame of reference, mainly based on literature studies, have also been an important part in the analysis that was executed. In general, it can be concluded that it is difficult to draw any conclusions about how handling of goods, delivery service and total cost of ownership would be affected by the implementation of a more coordinated freight distribution. Regarding the handling of goods, it is believed that the time spent handling goods will be reduced and that the amount of storage needed will increase, within the boundaries of the studied system. Since the information regarding delivery service in the outer flow is very limited, it is hard to draw any conclusions regarding the potential future changes to the delivery service with a greater inner flow.  The results of the total cost of ownership analysis indicates decreased costs. However, the different elements of the model cannot be compared to one another, and since the model is based on qualitative reasoning and empirical studies collected during the course of two weeks, the results must be viewed with caution. In conclusion, the study indicates that a more coordinated freight distribution could be a good idea. However, this should be investigated further, and data should be collected during a longer period of time. Furthermore, tangible plans for what a more coordinated freight distribution would look like in Region Östergötland, should be designed. / Region Östergötland ansvarar för leveranser av förbrukningsmaterial till sjukhus, vårdcentraler, folktandvårdskliniker och övriga kunder i regionen. Detta sker i nuläget genom två olika material-flöden, det inre respektive det yttre flödet. Leveranser från det inre flödet levereras till kund via det centrallager som ligger i Norrköping medan leveranser i det yttre flödet går direkt från extern leverantör till kund. Regionen vill idag få en bättre kontroll över flödet av förbrukningsmaterial och vill undersöka hur en mer samordnad varudistribution, och därmed ett ökat inre flöde, skulle påverka verksamheten sett ur ett helhetsperspektiv.  Syftet med detta arbete var således att studera hur hantering, leveransservice och totalkostnad skulle komma att påverkas vid införandet av en mer samordnad varudistribution av förbrukningsmaterial till vården i Region Östergötland. Detta för att ge beslutsunderlag för framtida beslut om materialförsörjningen inom regionen. De slutsatser som dragits bygger till stor del på den empiri som samlats in under arbetets gång. Empirin har samlats in på flera olika sätt och har varit både kvalitativ och kvantitativ. Intervjuer har genomförts med regionens anställda likväl som med anställda på några av de företag regionen har skrivna avtal med. Vidare så har både leveransprotokoll och enkäter skickats ut till regionens kunder och data från regionens affärssystem och nuvarande avtal har granskats. Rapportens referensram, till stor del baserad på litteraturstudier, har även den varit en viktig del i den analys som genomförts. Generellt kan det konstateras att det är svårt att dra slutsatser kring hur hantering, leveransservice och totalkostnad kommer att påverkas av en i större grad samordnad varudistribution. Hanterings-mässigt tros tid kunna sparas in sett ur ett helhetsperspektiv medan mer utrymme troligtvis kommer att krävas. Då informationen kring leveransservicen i det yttre flödet är väldigt knapphändig är det svårt att dra slutsatser kring hur leveransservicen ut till kund skulle påverkas av en potentiell förändring. Sett till genomförd totalkostnadsanalys indikeras minskade kostnader, men då de olika totalkostnadselementen är svåra att jämföra med varandra och då resultaten till stor del bygger på kvalitativa resonemang och empiri insamlad under två veckor bör resultaten beaktas med försiktighet. Sammanfattningsvis tyder studien på att en mer samordnad varudistribution skulle kunna vara en god idé. Detta bör dock undersökas närmre och data samlas in under en längre tidsperiod. Vidare bör konkreta planer utformas för hur en samordnad varudistribution skulle se ut i regionen.

The less-than-truckload (LTL) industry is highly competitive, with recent average profit margins less than 3%. LTL shipments are routed through a network of service centers and hubs. The performance of the entire LTL distribution network is highly dependent on the speed and accuracy of the hub operations. The focus of this research effort is to improve hub operations in order to reduce costs and increase service performance levels. Specifically, new approaches are investigated for assigning trailers to dock doors and sequencing the unloading of shipments at hubs. This thesis reviews current industry practices and available research literature on hub operations. Solution approaches for the trailer-to-door assignment and freight sequencing problems are presented along with case study results. The main performance measures are bottleneck time, total labor time, and total travel distance. For the trailer-to-door assignment problem, also referred to as the hub layout problem, the three approaches investigated are the original approach, a semi-permanent approach, and a dynamic approach. For the freight sequencing problem, the five approaches evaluated are trailer-at-a-time, trailer-at-a-time with offloading, nearest neighbor within a group, nearest neighbor within a shared group, and nearest neighbor. The approaches are implemented in C++ and analyzed using data from a regional LTL carrier. The case study results indicate that the dynamic layout performs significantly better than the original and semi-permanent layout for total distance, total labor time, and bottleneck time. For total distance and total labor time, the dynamic layout with nearest neighbor sequencing is the preferred approach. For bottleneck time, the dynamic layout with trailer-at-a-time with offloading performs best, while the nearest neighbor sequencing approach performs almost as well. In general, the case study results indicate that a dynamic layout with either a trailer-at-a-time with offloading approach or a nearest neighbor approach offers the largest potential for improvement. The assumptions and results of the hub layout and freight sequencing approaches are further evaluated using a simulation model. The simulation model indicates that a dynamic layout with nearest neighbor sequencing offers the largest potential for improvement in a more realistic environment with probabilistic and dynamic events. The simulation results also indicate that the trailer-at-a-time with offloading approach may need to be modified to account for more realistic dock conditions. In summary, the approaches explored in this research offer significant opportunity to improve hub operations through reducing bottleneck time, total labor time, and total travel distance. / Master of Science

La distribution de marchandises est un secteur en constant développement et constitue un facteur économique important. Par contre, dans les villes, il contribue notamment aux problèmes de congestion, pollution, bruit et d'autres dérangements à la population des villes. Pour faire face à ces problèmes, une nouvelle discipline est née à la fin du XXe siècle, la " City Logistics ", qui a comme objectifs principaux la réduction de la congestion, la pollution et le bruit occasionné par le transport de marchandises en ville. Dans les dernières années, plusieurs études et expériences se sont développées en toute l'Europe, mais pour l'instant une politique commune en matière de logistique urbaine n'a pas encore été proposée par l'Union Européenne. En Italie, seulement certaines villes de petite taille ont expérimenté des politiques de " city logistics " avec succès, mais sans un lien entre elles. Nous observons que ces expériences utilisent des centres urbains de distribution de marchandises, ce qui peut se traduire en un système de transport à deux ou plus niveaux. Plusieurs études en recherche opérationnelle ont traité des problématiques liées à des systèmes à niveaux multiples pour la distribution de marchandise. Néanmoins, l'optimisation des coûts de transport est en générale réalisé en considérant chaque niveau indépendant des autres, ou en approximant les coûts du transport dans certains niveaux pour simplifier. Un autre problème est le manque d'une unification de la terminologie utilisée dans ces études, qui difficulte la recherche bibliographique. Le but de cette recherche est, d'un coté, proposer des lignes guide d'accion en matière de planification de la distribution urbaine de marchandises, en unifiant certains termes, et d'un autre coté présenter une famille de problèmes d'optimisation de routes des véhicules qui considère les systèmes à niveaux multiples dans son ensemble et pas comme une somme de systèmes indépendants. Dans un premier temps, nous présentons les principales expériences de " city logistics " en Italie, ainsi que des lignes d'action dans la planification des systèmes de distribution urbaine des marchandises qui puissent devenir opérationnels et efficients. Ensuite nous présentons les principales problématiques et limites de l'optimisation de systèmes de transports à niveaux multiples, en unifiant les concepts et la notation. Nous proposons une nouvelle famille de problèmes d'optimisation de routes de véhicules pour des systèmes à niveaux multiples, en détaillant le cas basique : le problème de routes de véhicules à deux niveaux. Nous proposons des modèles mathématiques pour ce problème et des résultats numériques pour illustrer les avantages et les limites de la modélisation de ces systèmes.

With an increasing urban population around the world, the need for urban freight distribution is constantly growing. Many cities face problems with traffic congestion, especially during eak hours in the morning and afternoon. At the same time, the roads are often nearly empty during nighttime. In some cities, like Stockholm, heavy haulage is not permitted to enter the city during the night. This means that carriers are forced to perform these deliveries during the day, which leads to inefficient distribution. To investigate the possibilities to use nighttime hours for deliveries, which could lead to a more efficient distribution system and increased sustainability of the city, a pilot project has been initiated by the City of Stockholm to test off-peak deliveries. Other similar trial projects have encountered difficulties with getting businesses to participate, and the incentives to shift to offpeak deliveries have been unclear. Therefore, this thesis aims to investigate what incentives there are for actors within a supply chain to make this shift and the key factors that enable the supply chain to benefit. The thesis is performed as a case study on the pilot project in Stockholm, and uses a business model perspective to analyze how value is created and captured throughout the supply chain. The results show that there is potential for supply chains to increase its total value through offpeak deliveries as a result of increased efficiency, improved delivery reliability and increased utilization of trucks. Pricing models and relative negotiation power between actors have a large effect on how the value is distributed. The main contributors to creating increased value are sufficient delivery volume, compatible processes, and full utilization of trucks. The possibility to use off-peak deliveries for marketing is relatively unexploited and could potentially create more value. / Med  en  ökande  befolkning  i  städer  världen  över  ökar  också  behovet  av  varudistribution  istadskärnor. Många städer har problem med trängsel i trafiken, i synnerhet under rusningstid på morgonen och eftermiddagen. Samtidigt är vägarna ofta nästan tomma på natten. I en del städer, som Stockholm, är det förbjudet att köra tung lastbil i innerstaden under natten. Detta betyder att distributörer är tvingade att utföra dessa leveranser under dagtid, vilket leder till ineffektiv distribution.För att undersöka möjligheterna att utnyttja nattens timmar till leveranser, vilket skulle kunna leda till ett mer effektivt distributionssystem och en ökad hållbarhet för staden, har ett pilotprojekt initierats av Stockholms Stad för att testa nattliga leveranser. Andra liknande projekt har stött på problem med att få företag att deltaga, och incitamenten för att flytta leveranser till nattetid har varit otydliga. Därför syftar detta arbete till att undersöka vilka incitament som finns för aktörer inom en försörjningskedja att göra detta byte och vilka huvudfaktorer som möjliggör att kedjan kan dra fördel av det.Arbetet utförs i form av en fallstudie på pilotprojektet i Stockholm, och använder ett affärsmodellsperspektiv för att analysera hur värde skapas och fångas inom kedjan. Resultaten visar att det finns potential för försörjningskedjor att öka sitt totala värde genom nattliga leveranser som ett resultat av ökad effektivitet, förbättrad leveranssäkerhet och ökad nyttjandegrad av lastbilen. Prismodeller och relativ förhandlingsmakt mellan aktörer har en stor påverkan på hur värdet fördelas. De faktorer som bidrar mest till ett ökat värde är tillräcklig leveransvolym, kompatibla processer och hög nyttjandegrad av lastbilar. Möjligheten att använda nattliga leveranser för marknadsföringssyfte är relativt outnyttjat och skulle potentiellt kunna skapa mer värde.