Cette thèse s'intéresse au problème de conception optimale des routes de départ et d'arrivée dans une zone terminale autour d'un aéroport. Cette conception prend en compte la configuration et l'environnement autour des aéroports, et les différentes contraintes sous-jacentes, notamment l'évitement des obstacles et la séparation des routes. Nous proposons une formulation mathématique conduisant à un problème d'optimisation combinatoire, ainsi que des méthodes de résolution ad hoc efficaces pour le problème. Pour la résolution du problème, nous procédons en deux étapes. Nous considérons d'abord la conception d'une route de longueur minimale évitant les obstacles, en utilisant la méthode de Branch and Bound (B&B). Ensuite, nous nous intéressons à la conception de plusieurs routes en assurant en plus la séparation des routes. Deux approches différentes sont appliquées : une méthode basée sur la méthode B&B pour construire les routes séquentiellement suivant un ordre fixé à l'avance, et une méthode de recuit simulé pour construire les routes simultanément. Les résultats sur un ensemble de problèmes tests (artificiels et réels) montrent l'efficacité de notre approche. / This thesis proposes a methodology for the optimization of departure and arrival routes in the Terminal Maneuvering Area (TMA). The design of these routes takes into account the configuration and environment around airports, and the related constraints, in particular the avoidance of obstacles and the separation between routes. We propose a mathematical formulation leading to a combinatorial optimization problem, as well as efficient ad hoc resolution methods for the problem. The problem is solved in two steps. First, we design an individual route avoiding obstacles with respect to minimum route length by using a Branch and Bound (B&B) method. Afterwards, the design of multiple routes is solved by two different approaches: a B&B-based approach (where routes are generated sequentially in a given order) and a Simulated Annealing approach (where routes are generated simultaneously). The simulation results of a set of (artificial and real) test problems show the efficiency of our approach.

Med fokus på distribution behandlar denna rapport planerandet och utförandet av ruttplaneringen hos fallföretaget Gerdas Bröd i Grästorp AB. Utifrån antagandet att resurserna är begränsade hos Gerdas så presenteras ruttförslag utifrån målfunktionerna fyllnadsgrad och minsta möjliga körsträcka. Syftet har aldrig varit att presentera några optimala rutter utan snarare visa hur man på ett rimligt, logiskt och metodiskt sätt kan arbeta med dessa frågor utan att ta särskilt stora resurser i anspråk. Genom datainsamling bestående av intervjuer, observationer och samtal med flera rollinnehavare inom verksamheten har denna fallstudie tagit form. Företagets nuläge och önskan om åtgärder kunde tack vare denna datainsamling identifieras och lägga grunden för vidare analys och rekommendationer. De begränsade resurserna kan i vissa avseenden förklara avsaknaden av planering och styrning inom verksamheten. Det har lett till att chaufförerna i stor omfattning får sköta sig själva i utförandet av inlastning och leverans. Som ett ansikte utåt för verksamheten sköter förarna i praktiken all distribution samtidigt som de hanterar en stor andel av den kontinuerliga kundkontakten. Utförandet av ruttförslag är därför i stor utsträckning till för att stötta dem i deras dagliga arbete. Efter analys och jämförelse mellan teori och empiri mynnar detta ut i en rad olika rekommenderade åtgärder för fallföretaget. Med den kraftigt varierande efterfrågan behöver man vara flexibla internt för att kunna parera mot eventuell över- och underbemanning. Tillgång till fler bilar och förare är därför en av flera åtgärder som rekommenderas. / With focus on distribution, this rapport touches the planning and excecution of routes at the business case Gerdas Bröd in Grästorp AB. From the assumption that the resources are limited at Gerdas this rapport will present route suggestions from the objective functions load factor and minimum mileage. The purpose has never been to present any optimized routes but rather show how to work with these things in a reasonable, logical and methodically way without claiming any remarkable amount of resources. Trough data collection of interviews, observations and dialogs with several employees with different roles within Gerdas, this rapport has developed into this final version. The present of the company and their desire of actions could, thanks to the data collection, be identified and be the foundation of further analysis and recommendation. The limited resources can in some ways explain the lack of planning and management within the distribution of the company. This has led to the drivers being on their own when it comes to the execution of the loading and transporting. As the public face of the company the drivers handles almost everything when it comes to distribution while they manage to handle a big majority of the continuous customer contact. The makings of the route suggestions are primarily to support them in their daily work. After analysing and comparing theory with the company empiricism this will end in a line of recommended actions for the business case. With the greatly varying demand you must be internal flexible to be able to respond to any under- or over staffing. The access to more trucks and drivers is one of several recommended actions.

A segurança nos voos de Veículos Aéreos Não Tripulados (VANTs) é uma importante questão e vem ganhando destaque devido a uma série de acidentes com tais aeronaves. O aumento do número de aeronaves no espaço aéreo e a autonomia cada vez maior para realizar missões estão entre outros elementos que merecem destaques. No entanto, pouca atenção tem sido dada a autonomia da aeronave em casos emergenciais [Contexto]. Nesse contexto, o desenvolvimento de algoritmos que efetuem o planejamento de rotas na ocorrência de situações críticas é fundamental para obter maior segurança aérea. Eventuais situações de insegurança podem estar relacionadas a uma falha nos equipamentos do veículo aéreo que impede a continuação da missão em curso pela aeronave [Lacuna]. A presente pesquisa avança o estado da arte considerando um conceito chamado In-Flight Awareness (IFA), que estabelece consciência situacional em VANTs, visando maior segurança de voo. Os estudos também avançam na proposição de modelos matemáticos que representem o estado da aeronave avariada, viabilizando o pouso emergencial e minimizando possíveis danos [Propósito]. Este trabalho utiliza técnicas de computação evolutiva como Algoritmos Genéticos (AG) e Algoritmos Genéticos Multi-Populacional (AGMP), além de uma Heurística Gulosa (HG) e um modelo de Programação Linear Inteira Mista (PLIM) no tratamento de falhas críticas juntamente com o conceito de IFA [Metodologia]. As soluções obtidas foram avaliadas através de experimentos offline usando os modelos matemáticos desenvolvidos, além de validadas em um simulador de voo e em um voo real. De forma geral, o AG e AGMP obtiveram resultados equivalentes, salvando o VANT em aproximadamente 89% dos mapas. A HG conseguiu trazer a aeronave até uma região bonificadora em 77% dos mapas dentro de um tempo computacional abaixo de 1 segundo. No modelo PLIM, o tempo gasto foi de cerca de quatro minutos já que garantia a otimalidade da solução encontrada. Devido ao seu elevado tempo computacional, uma estratégia evolvendo rotas pré-calculadas foi definida a partir do PLIM, mostrando-se bastante promissora. Nos experimentos envolvendo simulador de voo foram testadas diferentes condições de vento e se verificou que mesmo sobre tais condições os métodos desenvolvidos conseguiram efetuar o pouso com segurança [Resultado]. O trabalho apresentado colabora com a segurança de Veículos Aéreos Não Tripulados e com a proposta de modelos matemáticos que representem a aeronave em caso de situações críticas. Os métodos, de forma geral, mostraram-se promissores na resolução do problema de pouso emergencial já que trouxeram a aeronave com segurança até regiões interessantes ao pouso em um baixo tempo computacional. Isso foi atestado pelos resultados obtidos a partir das simulações offline, em simulador de voo e em voo real [Conclusão]. As principais contribuições do trabalho são: modelagem de regiões adequadas ao pouso, modelagem de falhas, arquitetura do sistema planejador de rotas e modelo linear para para pouso emergencial [Contribuição]. / The security involved in flights of Unmanned Aerial Vehicles (UAVs) is an important issue and is achieving prominence due to a number of accidents involving such aircraft. Other elements that deserve highlights are the increase in the number of aircraft in the airspace and autonomy to perform missions, however, little attention has been given to the autonomy of the aircraft in emergency cases [Context]. In this context, the development of algorithms that contribute significantly to the path planning in the event of critical situations is essential for more air traffic. Possible situations of insecurity may be related to a failure in the equipment of vehicle that prevents the continuation of the current mission by aircraft [Gap]. The research advances the state of the art considering a concept called In-Flight Awareness (IFA), which provides situational awareness in UAVs aiming at greater flight safety. Advances also in the developing of mathematical models that represent the state of the damaged aircraft, with the purpose to execute the emergency landing by minimizing damages [Purpose]. Thus, this work applies evolutionary computation techniques such as Genetic Algorithms (GA) and Multi-Population Genetic Algorithms (MPGA), as well as a Greedy Heuristic (GH) and a Mixed Integer Linear Programming (MILP) model to deal with critical situations along with the concept of IFA [Methodology]. The solutions obtained were evaluated through offline experiments using the developed mathematical models, which were validated in a flight simulator and a real-world flight. In General, the GA and MPGA reached similar results by saving the UAV in approximately 89% of the maps, while the GH was able to bring the aircraft to a bonus region for 77% of maps within a feasible computational time lower than 1 second. In the MILP model, the time spent was about four minutes since it guarantees optimality of the solution found. Due to such high computational time, a strategy involving nearby routes pre-calculated was defined from the MILP which was very promising. In experiments involving flight simulator, different wind conditions were tested and it was found that even under such conditions the methods developed have managed to execute the landing safely [Result]. The work presented collaborates with the safety of Unmanned Aerial Vehicles and with the proposal of mathematical models that represent the aircraft under critical situations. The methods, in general, were promising since they brought the aircraft to execute a safe landing within a low computational time as shown by offline simulations, flight simulator and real flight [Conclusion]. The main contributions are: fault modeling, system architecture planner routes and linear model for emergency landing. [Contribution].

This master’s thesis is dedicated to user’s movement tracking in public transport, as well as to optimize route planning. In the first part, author presents information related to graph theory and graph theory algorithms, including Breadth-First Search, Djikstra’s algorithm and A*. In the second part of this thesis, author describes the design of server and mobile application including description of routing process and how individual parts works.

The problem of route planning in air transport (in public transport in general ) is similar to the shortest path problem. The main differences are the time dependency of the input graph and the multicriterial aspect of the path costs . The aim of this thesis was to create a complex system that would be able to load elementary flights from database and then respond to user requests with combined journeys made from single flights . It was achieved using two state of the art algorithms , CSA and RAPTOR , which were adapted for the flight graph. The experiments which were run on real world data showed massive speedup of the algorithms when using the proposed optimisations . The whole system was also tested against an existing proprietary solution .

This master’s thesis is about optimal route planning for electric vehicles. The first, theoretical part of this work introduces the issue and describes several heuristic methods that have been used to address optimization tasks. A practical part of the thesis is based around software implementation of the methods described earlier. Finally, several experiments and comparisons of obtained results are carried out.

Der Aufenthaltsort eines mobilen Benutzers stellt eine wichtige Information für Anwendungen aus den Bereichen Mobile Computing, Wearable Computing oder Ubiquitous Computing dar. Ist ein mobiles Endgerät in der Lage, die aktuelle Position des Benutzers zu bestimmen, kann diese Information von der Anwendung berücksichtigt werden -- man spricht dabei allgemein von ortsbezogenen Anwendungen. Eng verknüpft mit dem Begriff der ortsbezogenen Anwendung ist der Begriff des ortsbezogenen Dienstes. Hierbei handelt es sich beispielsweise um einen Dienst, der Informationen über den aktuellen Standort übermittelt. Mittlerweile werden solche Dienste kommerziell eingesetzt und erlauben etwa, dass ein Reisender ein Hotel, eine Tankstelle oder eine Apotheke in der näheren Umgebung findet. Man erwartet, nicht zuletzt durch die Einführung von LTE, ein großes Potenzial ortsbezogener Anwendungen für die Zukunft. Das jährlich stattfindende Fachgespräch "Ortsbezogene Anwendungen und Dienste" der GI/ITG-Fachgruppe Kommunikation und Verteilte Systeme hat sich zum Ziel gesetzt, aktuelle Entwicklungen dieses Fachgebiets in einem breiten Teilnehmerkreis aus Industrie und Wissenschaft zu diskutieren. Der vorliegende Konferenzband fasst die Ergebnisse des neunten Fachgesprächs zusammen. / The location of a mobile user poses an important information for applications in the scope of Mobile Computung, Wearable Computing and Ubiquitous Computing. If a mobile device is able to determine the current location of its user, this information may be taken into account by an application. Such applications are called a location-based applications. Closely related to location-based applications are location-based services, which for example provides the user informations about his current location. Meanwhile such services are deployed commercially and enable travelers for example to find a hotel, a petrol station or a pharmacy in his vicinity. It is expected, not least because of the introduction of LTE, a great potential of locations-based applications in the future. The annual technical meeting "Location-based Applications and Services" of the GI/ITG specialized group "Communication and Dsitributed Systems" targets to discuss current evolutions in a broad group of participants assembling of industrial representatives and scientists. The present proceedings summarizes the result of the 9th annual meeting.

Autonomous underwater vehicles have become a popular countermeasure to naval mines. Saab’s AUV62-MR detects, locates and identifies mine-like objects through three phases. By extracting functionality from the AUV62-MR and placing it on a second vehicle, it is suggested that the second and third phases can be performed in parallel. This thesis investigates how to design the second vehicle so that the runtime of the mine reconnaissance process is minimized. A simulation framework is implemented to simulate the second and third phases of the mine reconnaissance process in order to test various design choices. The vehicle design choices in focus are the size and the route planning of the second vehicle. The route-planning algorithms investigated in this thesis are a nearest neighbour algorithm, a simulated annealing algorithm, an alternating algorithm, a genetic algorithm and a proposed Dubins simulated annealing algorithm. The algorithms are evaluated both in a static environment and in the simulation framework. Two different vehicle sizes are investigated, a small and a large, by evaluating their performances in the simulation framework. This thesis takes into account the limited travelling distance of the vehicle and implements a k-means clustering algorithm to help the route planner determine which mine-like objects can be scanned without exceeding the distance limit. The simulation framework is also used to evaluate whether parallel execution of the second and third phases outperforms the current sequential execution. The performance evaluation shows that a major reduction in runtime can be gained by performing the two phases in parallel. The Dubins simulated annealing algorithm on average produces the shortest paths and is considered the preferred route-planning algorithm according to the performance evaluation. It also indicates that a small vehicle size results in a reduced runtime compared to a larger vehicle.

Traditional shortest path algorithms can efficiently find the optimal paths in graphs using simple heuristics. However, formulating a simple heuristic is challenging under the road network setting since there are multiple factors to consider, such as road segment length, edge centrality, and speed limit. This study investigates how a neural network can learn to take these factors as inputs and yield a path given a pair of origin and destination. The research question is formulated as: Are neural networks applicable to real-time route planning tasks in a roadnetwork?. The proposed metric to evaluate the effectiveness of the neural network is arrival rate. The quality of generated paths is evaluated by time efficiency. The real-time performance of the model is also compared between pathfinding in dynamic and static graphs, using theabove metrics. A staggered approach is applied in progressing this investigation. The first step is to generate random graphs, which allows us to monitor the size and properties of the training graph without caring too many details in a road network. The next step is to determine, as a proof of concept, if a neural network can learn to traverse simple graphs with multiple strategies, given that road networks are in effect complex graphs. Finally, we scale up by including factors that might affect the pathfinding in real road networks. Overall, the training data is optimal paths in a graph generated by a shortest path algorithm. The model is then applied to new graphs to generate a path given a pair of origin and destination. The arrival rate and time efficiency are calculated and compared with that of the corresponding optimal path. Experimental results show that the effectiveness, i.e., arrival rate ofthe model is 90% and the path quality, i.e., time efficiency has a medianof 0.88 and a large variance. The experiment shows that the model has better performance in dynamic graphs than in static graphs. Overall, the answer to the research question is positive. However, there is still room to improve the effectiveness of the model and the paths generated by the model. This work shows that a neural network trained to make locally optimal choices can hardly give a globally optimal solution. We also show that our method, only making locally optimal choices, can adapt to dynamic graphs with little performance overhead. / Traditionella algoritmer för att hitta den kortaste vägen kan effektivt hitta de optimala vägarna i grafer med enkel heuristik. Att formulera en enkel heuristik är dock utmanande för vägnätverk eftersom det finns flera faktorer att överväga, såsom vägsegmentlängd, kantcentralitet och hastighetsbegränsningar. Denna studie undersöker hur ett neuralt nätverk kan lära sig att ta dessa faktorer som indata och finna en väg utifrån start- och slutpunkt. Forskningsfrågan är formulerad som: Är neuronnätverket tillämpliga på realtidsplaneringsuppgifter i ett vägnät?. Det föreslagna måttet för att utvärdera effektiviteten hos det neuronnätverket är ankomstgrad. Kvaliteten på genererade vägar utvärderas av tidseffektivitet. Prestandan hos modellen jämförs också mellan sökningen i dynamiska och statiska grafer, med hjälp av ovanstående mätvärden. Undersökningen bedrivs i flera steg. Det första steget är att generera slumpmässiga grafer, vilket gör det möjligt för oss att övervaka träningsdiagrammets storlek och egenskaper utan att ta hand om för många detaljer i ett vägnät. Nästa steg är att, som ett bevis på konceptet, undersöka om ett neuronnätverk kan lära sig att korsa enkla grafer med flera strategier, eftersom vägnätverk är i praktiken komplexa grafer. Slutligen skalas studien upp genom att inkludera faktorer som kan påverka sökningen i riktiga vägnät. Träningsdata utgörs av optimala vägar i en graf som genereras av en algoritm för att finna den kortaste vägen. Modellen appliceras sedan i nya grafer för att hitta en väg mellan start och slutpunkt. Ankomstgrad och tidseffektivitet beräknas och jämförs med den motsvarande optimala sökvägen. De experimentella resultaten visar att effektiviteten, dvs ankomstgraden av modellen är 90% och vägkvaliteten dvs tidseffektiviteten har en median på 0,88 och en stor varians. Experimentet visar att modellen har bättre prestanda i dynamiska grafer än i statiska grafer. Sammantaget är svaret på forskningsfrågan positivt. Det finns dock fortfarande utrymme att förbättra modellens effektivitet och de vägar som genereras av modellen. Detta arbete visar att ett neuronnätverk tränat för att göra lokalt optimala val knappast kan ge globalt optimal lösning. Vi visar också att vår metod, som bara gör lokalt optimala val, kan anpassa sig till dynamiska grafer med begränsad prestandaförlust.

In route planning the goal is to obtain the best route between a set of locations, which becomes a very complex task as the number of locations increase. This study will consider the problem of transfer bus route planning and examines the feasibility of applying a reinforcement learning method in this specific real-world context. In recent research, reinforcement learning methods have emerged as a promising alternative to classical optimization algorithms when solving similar problems. This due to their positive properties in terms of scalability and generalization. However, the majority of said research has been performed on strictly theoretical problems, not using real-world data. This study implements an existing reinforcement learning model and adapts it to fit the realms of transfer bus route planning. The model is trained to generate optimized routes in terms of time and cost consumption. Then, routes generated by the trained model are evaluated by comparing them to corresponding manually planned routes. The reinforcement learning model produces routes that outperforms manually planned routes with regards to both examined metrics. However, due to delimitations and assumptions made during the implementation, the explicit differences in consumptions are considered promising but cannot be taken as definite results. The main finding is the overarching behavior of the model, implying a proof of concept; reinforcement learning models are usable tools in the context of real-world transfer bus route planning. / Inom ruttplanering är målet att erhålla den bästa färdvägen mellan en uppsättning platser, vilket blir en mycket komplicerad uppgift i takt med att antalet platser ökar. Denna studie kommer att behandla problemet gällande ruttplanering av transferbussar och undersöker genomförbarheten av att tillämpa en förstärkningsinlärningsmetod på detta verkliga problem. I nutida forskning har förstärkningsinlärningsmetoder framträtt som ett lovande alternativ till klassiska optimeringsalgoritmer för lösandet av liknande problem. Detta på grund utav deras positiva egenskaper gällande skalbarhet och generalisering. Emellertid har majoriteten av den nämnda forskningen utförts på strikt teoretiska problem. Denna studie implementerar en befintlig förstärkningsinlärningsmodell och anpassar den till att passa problemet med ruttplanering av transferbussar. Modellen tränas för att generera optimerade rutter, gällande tids- och kostnadskonsumtion. Därefter utvärderas rutterna, som genererats av den tränade modellen, mot motsvarande  manuellt planerade rutter. Förstärkningsinlärningsmodellen producerar rutter som överträffar de manuellt planerade rutterna med avseende på de båda undersökta mätvärdena. På grund av avgränsningar och antagandet som gjorts under implementeringen anses emellertid de explicita konsumtionsskillnaderna vara lovande men kan inte ses som definitiva resultat. Huvudfyndet är modellens övergripande beteende, vilket antyder en konceptvalidering; förstärkningsinlärningsmodeller är användbara som verktyg i sammanhanget gällande verklig ruttplanering av transferbussar.