La capacité à gagner du temps et à diminuer ses efforts est l'une des qualités de l'être humain, qui a conduit à exercer la pensée depuis l'Antiquité jusqu'à ces dernières décennies, caractérisées par l'émergence du mélange entre la rapidité des calculs et la précision des résultats, et ce dans plusieurs domaines. Le problème des tournées de véhicules et ses extensions sont, pour les théoriciens de ces utilités, d'une réelle importance quant aux applications du monde réel. Des recherches récentes dans ce domaine ont permis des avancées significatives dans la formulation des problèmes ainsi que dans la conception et l'analyse d'algorithmes. Dans cette étude, nous nous intéressons au problème de la logistique. Notre attention se porte en particulier sur un cas des réseaux de télécommunication, 2ECON-NDPR, et sur la façon de créer des designs d'une manière intelligente pour assurer la vitalité et la durabilité de la circulation de l'information. En outre. Nous choisissons les variantes problème de tournées de véhicules avec fenêtres de temps et problème de tournées de véhicules sélectives des familles VRP et OP respectivement. C'est dans ce cadre que s'inscrit cette thèse. La conception des solutions pour ces problèmes fait appel à la technique de programmation approchée connue pour sa rapidité de calcul. Il s’agit de Beam-search et de la recherche locale à grand voisinage. Nous présentons tout d’abord une étude détaillée des dernières problématiques précitées ainsi que différents types de méthodes de résolutions. Puis, nous exposons une méthode de recherche locale à grand voisinage adaptée pour la conception de réseau de survie avec relais, une proposition d’un algorithme de résolution approchée à trois phases pour le CVRPTW et, enfin, une proposition d'un algorithme de résolution approchée hybride pour le TOP. / The need to save time as well as minimize effort is part of the human condition and it has driven our though s from antiquity until these last few decades, now characterized by the emergence of a mix in all fields between rapidity of calculation and precision in the result. The vehicle routing problem and its extensions are an important field for theorists of these utilities for real-world applications. Recent research in the field has led to significant advantages in problem formulation and designing algorithm analyses. This study considers logistics problems. A particular locus was given to a certain case of telecommunications networks 2ECONNDPR, as well as the method of intelligently creating designs to ensure vitality and durability in information circulation. Furthermore, the study considered vehicle routing problems, with time windows and orienteering problems from the VRP and OP families, respectively. This is the framework for this thesis. Solutions to these problems use programming techniques known for their calculation speed, i .e ., Beam-search and very large-scale neighborhood searching. First, a detailed study is presented of these above mentioned problems, along with the various types or resolution methods. Next, a very large-scale neighborhood search method is presented, suited to the design of a survivable network with relay, a proposition for a three-stage heuristic for the capacitated vehicle routing problem with time windows and, finally, a proposition for a hybrid heuristic for the team orienteering problem.

This master thesis describes early attempts at using electroglottography (EGG) to capture such changes in vocal fold vibration patterns that could be of interest to teachers of contemporary commercial music. After initial explorations, focus is placed on detecting potentially detrimental phonation; more specifically on the pressed quality often associated with loud singing in high register (belting). FonaDyn, a program written in the SuperCollider language, is used to detect pressedness using an algorithm based on K-means clustering of Fourier components of EGG cycles. Results indicate that pressedness affects phonation in ways detectable using EGG. Changes caused by pressedness seem to vary between registers and this variation is similar between subjects. Detection of pressedness in a subject is quite successful when training the algorithm on the same subject, but not always across subjects. / Denna masteruppsats beskriver inledande försök att använda elektroglottografi (EGG) för att avläsa sådana förändringar i stämbandens vibrationsmönster som skulle kunna vara av intresse för sånglärare inom icke-klassisk stil. Tidiga undersökningar leder till att fortsatt fokus läggs på att detektera fonationstyper som kan orsaka röstskador; mer specifikt den typ av pressad röstkvalitet som ofta förknippas med stark sång i högt register (s.k. belting). FonaDyn, ett datorprogram skrivet i språket SuperCollider, används för att detektera pressad fonation med hjälp av K-means-klustring av EGG-cykler baserat på deras Fourierkomponenter. Resultaten indikerar att pressad fonation går att urskilja med hjälp av EGG. Kännetecknen för pressad fonation tycks skilja sig mellan röstregister och denna skillnad är snarlik hos olika försökspersoner. Programmet klarar av att känna igen pressad fonation hos samma person som algoritmen tränats på men inte alltid om algoritmen tränats på en annan sångare.

Examensarbetet handlar om att minimera en total servicekostnad för företaget Chargestorm då de skall utföra service på sina laddningscentraler för elbilar. En ruttoptimering har gjorts som resulterar i ett antal rutter som är tänka att användas av en servicetekniker när denne åker runt och utför service på laddningscentralerna. En simulering har också gjorts för att påvisa de bästa tidsintervallen då olika typer av service kan ske, med målet att minimera den totala servicekostnaden.

One of the significant side-effects of growing urbanization is the constantly increasing amount of freight transportation in cities. This is mainly performed by conventional vans and trucks and causes a variety of problems such as road congestion, noise nuisance and pollution. Yet delivering goods to residents is a necessity. Sustainable concepts of city distribution networks are one way of mitigating difficulties of freight services. In this paper we develop a two-echelon city distribution scheme with temporal and spatial synchronization between cargo bikes and vans. The resulting heuristic is based on a greedy randomized adaptive search procedure with path relinking. In our computational experiments we use artificial data as well as real-world data of the city of Vienna. Furthermore we compare three distribution policies. The results show the costs caused by temporal synchronization and can give companies decision-support in planning a sustainable city distribution concept.

In this work we present a multiagent Fleet Coordination Problem (FCP). In this formulation, agents seek to minimize the fuel consumed to complete all deliveries while maintaining acceptable on-time delivery performance. Individual vehicles must both (i) bid on the rights to deliver a load of goods from origin to destination in a distributed, cooperative auction and (ii) choose the rate of travel between customer locations. We create two populations of adaptive agents, each to address one of these necessary functions. By training each agent population in separate source domains, we use transfer learning to boost initial performance in the target FCP. This boost removes the need for 300 generations of agent training in the target FCP, though the source problem computation time was less than the computation time for 5 generations in the FCP. / Graduation date: 2012

This thesis identifies some problems, the genetic algorithm (GA) is facing in the area of vehicle routing and proposes various methods to address those problems. Those problems arise from the unavailability of suitable chromosomal representation and evaluation schemes of GA for the Vehicle Routing Problem (VRP). The representation and evaluation schemes already in use have problems of high computational cost, illegal chromosomes (chromosomes not representing a legal tour) and wrong fitness assignment (fitness not truly representing chromosome genetic makeup). These problems are addressed by several proposed new schemes, namely the Self Imposed Constraints Evaluation scheme, the Contour and Reverse Contour Evaluation schemes and the Order Skipping Evaluation scheme, which are specifically tailored for various objectives, problems and situations. Apart from this, a methodology, which has previously being used in other meta-heuristics, is incorporated into GA i.e., the independent application of GA on various sub-localities of the problem. We call this GA, a Localized Genetic Algorithm (LGA). LGA is an iterative procedure between optimization and controlled de-optimization. The procedure of controlled de-optimization is also novel. It brings the solution into a new search space while controlling its cost effectively. LGA is introduced with various search techniques, i.e. intensive, extensive and selective, the use of which depends on the problem size and the availability of computational resources. Furthermore, search reduction techniques (Fitness Approximation Methods) are also introduced into the LGA, which has enabled the LGA to be applied to large scale problems. Due to the implementation of those proposals, LGA is the first GA-driven approach to be applied to very large scale CVRP problems of up to 1200 customers, i.e. datasets presented by Feiyue in 2005 and large scale VRPTW problems of up to 1000 customers, datasets presented by Gehring and Homberger in 1999. Lastly, a standard unit for computational comparison, i.e., Bellman's Evaluation Units BEUs, is also introduced to facilitate computational comparisons for future researchers. LGA has shown promising results on CVRP and VRPTW problems. It is flexible and also has the potential to be extended to not only other vehicle routing problems, but also to other ordering problems.

Prompt and efficient intervention is vital in reducing casualty figures during epidemic outbreaks, disasters, sudden civil strife or terrorism attacks. This can only be achieved if there is a fit-for-purpose and location-specific emergency response plan in place, incorporating geographical, time and vehicular capacity constraints. In this research, a comprehensive emergency response model for situations of uncertainties (in locations' demand and available resources), typically obtainable in low-resource countries, is designed. It involves the development of algorithms for optimizing pre-and post-disaster activities. The studies result in the development of four models: (1) an adaptation of a machine learning clustering algorithm, for pre-positioning depots and emergency operation centers, which optimizes the placement of these depots, such that the largest geographical location is covered, and the maximum number of individuals reached, with minimal facility cost; (2) an optimization algorithm for routing relief distribution, using heterogenous fleets of vehicle, with considerations for uncertainties in humanitarian supplies; (3) a genetic algorithm-based route improvement model; and (4) a model for integrating possible new locations into the routing network, in real-time, using emergency severity ranking, with a high priority on the most-vulnerable population. The clustering approach to solving dept location-allocation problem produces a better time complexity, and the benchmarking of the routing algorithm with existing approaches, results in competitive outcomes.

The master's thesis deals with the new modification of vehicle routing problem -- 3PL vehicle routing problem with heterogeneous fleet and split delivery. In addition to classical vehicle routing problem, we consider a heterogeneous suppliers fleet and also external carrier, which charges a fixed value per unit of transported goods. The reader is first introduces to vehicle routing problem, its history and possible solutions. Furthermore, the reader is acquainted with logistics and logistics providers. In the main part of this thesis is described 3PL vehicle routing problem and its mathematical model. At first we try to get optimal solution via CPLEX solver. But since this is an NP-hard task, heuristic method is proposed (in two variants) for solving this problem. The heuristic is then tested on the selected test tasks. Results obtained using the proposed heuristics are compared with the optimal solution. Even larger problems are then solved using this heuristics. In the end other modifications and possible improvements of this heuristic method are proposed.

En traditionell sånguppvärmning består ofta av att man fonerar (bildar ljud) olika frekvenser (Hz) med varierande ljudstyrka (dB SPL). I viss mån utförs även uppvärmningsövningar såsom stretch av nacke och olika kroppsliga rörelser för att öka blodcirkulationen. En del sångpedagoger förespråkar träning före sång. Några fördelar med fysisk träning före sång kan vara ökat luftflöde, sänkt ”Phonation Threshold Pressure” (PTP) och högre tonsäkerhet. Det finns även studier som visar fördelar med en traditionell sånguppvärmning däribland högre tonsäkerhet och en sänkning av PTP. I denna uppsats undersöks hur fysisk träning fungerar som sånguppvärmning och att jämföra detta med en traditionell sånguppvärmning. Fyra testpersoner rekryterades till studien via personliga kontakter, samtliga män och i 30-årsåldern. Personerna utförde två tester vid två olika tillfällen vardera enligt en counterbalance strategi, det vill säga att två av personerna utförde en traditionell uppvärmning vid det första testtillfället och de andra två utförde en fysisk uppvärmning vid det första testtillfället. Efter att ha utfört en uppvärmning fick de sjunga en sång, utföra en ”Voice range profile” (VRP) samt skatta sina egna upplevelser av testtillfället via enkät. Eftersom det ingick fyra personer i undersökningen går det inte att dra några generella slutsatser men i studien sjöng alla med svagare ljudstyrka (dB SPL) efter utförandet av den fysiska uppvärmningen. Detta kan tyda på en sänkning av PTP. Alla testpersoner hade även ett större frekvensomfång (Hz) efter den fysiska uppvärmningen. Tre av fyra sjöng med högre amplitud (dB SPL) och hade även ett större omfång gällande dB SPL efter den traditionella uppvärmningen. Föreliggande studie visar på fördelar med både den fysiska uppvärmningen och den traditionella. Den som svarade bäst på den fysiska uppvärmningen var den mest erfarne sångaren.

This dissertation focuses on supply chain design and logistics problems with emphasis on biomass logistics and routing problems. In biomass logistics, we have studied problems arising in a switchgrass-based bio-ethanol supply chain encountered in the Southeast, and a corn stover harvest scheduling problem faced in the Midwest Unites States, both pertaining to the production of cellulosic ethanol. The main contributions of our work have been in introducing new problems to the literature that lie at the interface of the lot-sizing and routing problems, and in developing effective exact algorithms for their solution. In the routing area, we have addressed extensions of the well-known traveling salesman and vehicle routing problems. We have proposed new formulations and have developed exact algorithms for the single and multiple asymmetric traveling salesmen problems (ATSP and mATP), the high-multiplicity asymmetric traveling salesman problem (HMATSP) and its extensions, and the fixed-destination multi-depot traveling salesman problem with load balancing (FD-MTSPB). Furthermore, we have introduced a new strategy to reduce routing cost in the classical vehicle routing problem (VRP). / Ph. D.