In real world applications such as the transporting of goods products, packing is a major issue. Goods products need to be packed such that the smallest space is wasted to achieve the maximum transportation efficiency. Packing becomes more challenging and complex when the product is circular/spherical. This thesis focuses on the best way to pack three-dimensional unit spheres into the smallest spherical and cubical space. Unit spheres are considered in lieu of non-identical spheres because the search mechanisms are more difficult in the latter set up and any improvements will be due to the search mechanism not to the ordering of the spheres. The two-unit sphere packing problems are solved by approximately using a variable neighborhood search (VNS) hybrid heuristic. A general search framework belonging to the Artificial Intelligence domain, the VNS offers a diversification of the search space by changing neighborhood structures and intensification by thoroughly investigating each neighborhood. It is exible, easy to implement, adaptable to both continuous and discrete optimization problems and has been use to solve a variety of problems including large-sized real-life problems. Its runtime is usually lower than other meta heuristic techniques. A tutorial on the VNS and its variants along with recent applications and areas of applicability of each variant. Subsequently, this thesis considers several variations of VNS heuristics for the two problems at hand, discusses their individual efficiencies and effectiveness, their convergence rates and studies their robustness. It highlights the importance of the hybridization which yields near global optima with high precision and accuracy, improving many best- known solutions indicate matching some, and improving the precision and accuracy of others. Keywords: variable neighborhood search, sphere packing, three-dimensional packing, meta heuristic, hybrid heuristics, multiple start heuristics.

Orientadores: Sandra Augusta Santos, Jose Mario Martinez / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Computação Cientifica / Made available in DSpace on 2018-08-08T17:04:51Z (GMT). No. of bitstreams: 1 Xavier_LarissaOliveira_M.pdf: 1818711 bytes, checksum: e0305d93bde788c50b64809da1b8bf9e (MD5) Previous issue date: 2007 / Resumo: Este trabalho encaminha a investigação de questões relacionadas ao desempenho de algoritmos de região de confiança para problemas de otimização irrestrita de grande porte. O algoritmo clássico de Moré e Sorensen, baseado em fatorações de Cholesky, é comparado com a abordagem de Rojas, Santos e Sorensen (algoritmo RSS). Do ponto de vista teórico são estudados os resultados de convergência dos dois algoritmos. Em termos práticos, são resolvidos problemas com a estrutura típica de empacotamento de cilindros. Também são pesquisados o desempenho efetivo do algoritmo RSS na solução aproximada dos subproblemas, e a repercussão da precisão com que os subproblemas são resolvidos no esforço global do algoritmo. / Abstract: This work investigates issues related to the performance of trust-region algorithms for large-scale unconstrained minimization. The classic algorithm of Moré and Sorensen, based on Cholesky?s factorizations, is compared with the approach of Rojas, Santos and Sorensen (algorithm RSS). From the theoretic standpoint, the convergence results of both algorithms are compiled. In practical terms, problems with the typical structure of packying of cylinders are solved. The effective performance of the algorithm RSS in the approximate solution of the subproblems is analyzed as well, together with the influence of the inner precision of the subproblems to the global effort of the algorithm / Mestrado / Otimização / Mestre em Matemática Aplicada

Orientador: José Mario Martínez Pérez / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Matemática, Estatística e Computação Científica / Made available in DSpace on 2018-08-20T05:08:42Z (GMT). No. of bitstreams: 1 FloresCallisaya_Hector_D.pdf: 2324904 bytes, checksum: e15e7624ccad0fdf64ce3c4d8095c20a (MD5) Previous issue date: 2012 / Resumo: Neste trabalho, serão propostos modelos matemáticos para problemas de empacotamento não reticulado de esferas em regiões limitadas por quadráticas no plano e no espaço. Uma técnica para construir representações ou parametrizações será introduzida, mediante a qual será possível encontrar um sistema de desigualdades que determinam o empacotamento de um número fixo de esferas. Desta forma, resolvemos o problema de empacotamento de esferas através de uma sequência de sistemas de desigualdades. Finalmente, para obter resultados eficientes, minimizaremos a função de sobreposição, usando o método do Lagrangiano Aumentado / Abstract: In this work, we will propose mathematical models for not latticed packing of spheres problems in regions bounded by quadratic in the plane and in the space. A technique to construct representations or parameterizations will be introduced, by which it will be possible to find a system of inequalities which determine the packing of a fixed number of spheres. Thus, we solve the problem of packing spheres through a sequence of systems of inequalities. Finally, to obtain effective results, we will minimize the overlay function using the Augmented Lagrangian Method / Doutorado / Matematica Aplicada / Doutor em Matemática Aplicada