There are many times in which a critical choice between proposed system architectures must be made. Two situations in particular motivate this dissertation: a "Cambrian explosion" when no dominant rchitecture has arisen, and times in which developments enable challenges to a dominant incumbent. In each situation, the advance of core technologies is key. This dissertation features a new computing technique to systematically explore the interaction of technological progress with architectural choices. This technique is founded upon a graph theoretic formulation of architecture, which enables the consideration of multifunctional components and modularity v. synergy trades. The technique utilizes a genetic algorithm formulated for graphs, and a solver that automatically constrains and optimizes component design variables. The use of quantitative technology models, graph theoretic formulation, and optimization algorithms together enables a systematic exploration of both time and combinatorial spaces. The quantitative results of this exploration enhance the strategic view of technology planners.

Thesis (Ph. D.)--State University of New York at Binghamton, Department of Computer Science, Thomas J. Watson School of Engineering and Applied Science, 2007. / Includes bibliographical references.

Thesis (Ph. D.)--Mathematics, Georgia Institute of Technology, 2008. / Committee Chair: Thomas, Robin; Committee Co-Chair: Cook, William J.; Committee Member: Dvorak, Zdenek; Committee Member: Parker, Robert G.; Committee Member: Yu, Xingxing.

Thesis (M.S.)--University of Texas at El Paso, 2008. / Title from title screen. Vita. CD-ROM. Includes bibliographical references. Also available online.

Thesis (M.S. in Mechanical Engineering)--Naval Postgraduate School, March 2005. / Thesis Advisor(s): Fotis Papoulias, Roman B. Statnikov. Includes bibliographical references (p. 241). Also available online.

Thesis (Ph.D.)--Civil and Environmental Engineering, Georgia Institute of Technology, 2009. / Committee Chair: Aral, Mustafa; Committee Member: Guan, Jiabao; Committee Member: Kim, Seong-Hee; Committee Member: Roberts, Philip; Committee Member: Uzer, Turgay.

Thesis (Ph.D.) - Carleton University, 2007. / Includes bibliographical references (p. 115-129). Also available in electronic format on the Internet.

Dans le cadre de cette thèse, nous nous intéressons au traitement des problèmes d’optimisation combinatoire liés à la conception d’outils de gestion des systèmes de véhicules partagés. Ces problèmes sont proches des problèmes de collecte et de livraison. Après avoir réalisé une étude théorique sur des problèmes d’optimisation combinatoire autour du transport et des méthodes de résolutions, nous nous sommes intéressés ici à trois problèmes particuliers : le PPRV, le PPRV-PM et le PPRV-T. Le premier problème est le Problème de Planification du Redéploiement de Véhicules partagés (PPRV). C’est une extension du One-commodity Pickup-and-Delivery Problem (1-PDP) car les véhicules partagés sont indifférenciés. Nous avons proposé un modèle linéaire et une heuristique utilisant le schéma hybride ILS/VND. L’approche développée repose sur la stratégie « route-first, cluster-second » : on commence par construire une tournée géante, puis on l’améliore par une procédure de perturbation et une recherche locale. Pendant la recherche locale, la contrainte de capacité des véhicules est momentanément relaxée et progressivement restaurée ; la tournée géante obtenue est ensuite transformée en plusieurs tournées à l’aide de la procédure Split. Les deux problèmes suivants sont considérés comme des extensions du PPRV en autorisant des livraisons partielles : PPRV avec Passage Multiple (PPRV-PM) et PPRV avec Transfert d’objets (PPRV-T). Nous proposons une approche de type « divide-first, route-second » pour la résolution du PPRV-PM. Elle consiste à effectuer d’abord un fractionnement de la demande, puis la résoudre à l’aide d’un schéma hybride de type GRASP/VND. Le PPRV-T étend le PPRV-PM au transfert d’objets entre les transporteurs lors du passage sur un sommet. Nous avons reformulé le PPRV-T comme un problème de multi-flots couplés sur un réseau dynamique. Nous avons proposé une méthode d’insertion basée sur cette modélisation. / In this thesis, we are interested to deal with combinatorial optimization problems related to design management tools for vehicle-sharing systems. These problems are close to the Pickup-and-Delivery Problems (PDP) in the literature. After performing a survey on the problems area and on the resolution methods, we focused on three specific problems and we proposed one approach for each problem. The first one is the sharing Vehicles Redeployment Planning Problem (VRPP), which is considered as a multi-vehicles extension of the One-commodity Pickup-and-Delivery Problem (1-PDP). We proposed a linear model and a hybrid heuristic which combines the ILS and VND. The proposed approach uses the rout-first, cluster-second strategy: we construct a Hamiltonian route, and then improve it using a procedure combines a shacking step and a VND local search. The used neighborhoods are adapted to the relaxation of capacity; the obtained route would be then split into several vehicles tours in the clustering phase.The two following problems are considered as extensions of VRPP introducing the split demand constraint : VRPP with Multi-Passage (VRPP-MP) and VRPP with Transferring objects (VRPP-T). We proposed an approach with the divide-first, route-second strategy for VRPP-MP. It consists of dividing in advance the demand, and then solves it using a hybrid scheme of GRASP/VND. In the VRPP-T, the objects carried could be exchanged between carriers when crossing on the sites. The VRPP-T is modeled here as a multi-flows problem on a dynamic network. We proposed an insertion method based on this modeling.

[pt] Esta tese define os problemas das p-medianas conectadas e o de localização de facilidades não-capacitadas conectadas. Possíveis aplicações incluem problemas de planejamento regional e o projeto de redes de telecomunicações ou de transporte. Para o primeiro problema, duas formulações de programação linear inteira são apresentadas e comparadas. Um destes modelos é adaptado para o segundo problema. Para o problema das p-medianas conectadas, algoritmos aproximados são desenvolvidos. Uma estratégia de busca local híbrida é proposta. Para acelerar as iterações do algoritmo de busca local, idéias como circularidade, melhoria iterativa e o descarte de vizinhos são incorporadas. Heurísticas GRASP e VNS são desenvolvidas incluindo a utilização de um filtro com o objetivo de diminuir os tempos de processamento e do procedimento de reconexão por caminhos com o objetivo de melhorar a qualidade das soluções encontradas. Diversos testes são realizados comparando-se esses algoritmos. Os resultados mostraram a necessidade de se executar um passo adicional de pós-otimização às heurísticas GRASP e VNS propostas. / [en] In this work, the connected p-median and the connected facility location problems are defined. Applications arise in regional planning, design of telecommunications and transportation networks. For the first problem, two integer linear programming formulations are proposed. Adaptations are made in one of these formulations and are used to model the second problem. Approximation algorithms to solve the connected p-median problem are developed. A hybrid local search strategy is proposed. In order to speed up the local search iterations, ideas as circularity, first- improving strategy and discard neighbors are incorporated. A GRASP algorithm and a VNS heuristic are also proposed. A filter is used to reduce the computational time required and a path-relinking is applied to improve the results found. Computational experiments to compare the algorithms are reported. To improve these results, it is applied a post-optimization step to the GRASP and VNS heuristics.

DANTAS, Allberson Bruno de Oliveira. Uma interface de programação distribuída para aplicações em otimização combinatória. 2011. 86 f. Dissertação (Mestrado em ciência da computação)- Universidade Federal do Ceará, Fortaleza-CE, 2011. / Submitted by Elineudson Ribeiro (elineudsonr@gmail.com) on 2016-07-08T17:57:51Z No. of bitstreams: 1 2011_dis_abodantas.pdf: 805347 bytes, checksum: c9671608a7d738f843239856e546e201 (MD5) / Approved for entry into archive by Rocilda Sales (rocilda@ufc.br) on 2016-07-13T12:23:03Z (GMT) No. of bitstreams: 1 2011_dis_abodantas.pdf: 805347 bytes, checksum: c9671608a7d738f843239856e546e201 (MD5) / Made available in DSpace on 2016-07-13T12:23:03Z (GMT). No. of bitstreams: 1 2011_dis_abodantas.pdf: 805347 bytes, checksum: c9671608a7d738f843239856e546e201 (MD5) Previous issue date: 2011 / This work was motivated by the need of exploiting the potential of distributed paralelism in combinatorial optimization applications. propose a distributed programming interface, To achieve this goal, we in which we cherish two main requirements: e ciency and reuse. The rst stems from the need of HPC (High applications require maximum possible performance. Performance Computing) Therefore, we specify our interface as an extension of the MPI library, which is assumed to be e cient for distributed applications. The reuse requirement must make compatible two important features: asynchronism and collective operations. Asynchronism must be present at our interface, once most of combinatorial optimization applications have an asynchronous nature. Collective operations are features that should be available in the interface, so that they can be used by applications in their execution. In order reach the reuse requirement, we based this interface on the Event- and Pulse-driven Models of Distributed Computing, once they are asynchronous and allow the incorporation of collective operations. We implemented partially the interface de ned in this work. In order to validate the use of the inteface by combinatorial optimization applications, we selected two applications and implemented them using our interface. They are the Branch-and-Bound technique and the Maximum Stable Set Problem (MSSP). We also provide some experimental results. / Este trabalho foi motivado pela necessidade da exploração do potencial do paralelismo distribuído em aplicações em Otimização Combinatória. Para tanto, propomos uma interface de programação distribuída, na qual prezamos dois requisitos principais: eficiência e reuso. O primeiro advém da necessidade de aplicações de CAD exigirem máximo desempenho possível. Assim sendo, especificamos esta interface como uma extensão da biblioteca MPI, a qual é assumida como eficiente para aplicações distribuídas. O requisito reuso deve tornar compatíveis duas características importantes: assincronismo e operações coletivas. O assincronismo deve estar presente na interface, uma vez que as aplicações em Otimização Combinatória, em sua maioria, possuem uma natureza assíncrona. Operações coletivas são funcionalidades que devem estar disponíveis na interface, de modo que possam ser utilizadas por aplicações em suas execuções. Tendo em vista atender o requisito reuso, baseamos esta interface nos Modelos de Computação Distribuída Dirigidos por Eventos e por Pulsos, pois os mesmos são assíncronos e permitem a incorporação de operações coletivas. Implementamos parcialmente a inteface definida neste trabalho. Tendo em vista validar uso desta inteface por aplicações em Otimização Combinatória, selecionamos duas aplicações e as implementamos utilizando a interface. São elas a técnica Branch-and-Bound e o Problema do Conjunto Independente Máximo (CIM). Fornecemos também alguns resultados experimentais.