The principal aim of the project has been to investigate formulation-solution processes and the extent to which these processes lead to better guidance and understanding of teaching, learning, and assessment in mathematical modelling. The following main activities have been carried out in support of this aim: the development of case studies of the mathematical modelling approaches that may be used in the solution of practical problems; the design of teaching and learning experiments carried out mainly with undergraduates with some knowledge of physics and teachers on an MSc course in mathematical education; the theoretical development of formulation-solution processes by means of a concept matrix and a relationship level graph; the analysis of a selection of students' modelling attempts; an investigation of assessment methods and the implications of the theoretical development of formulation solution processes for these methods. The case studies were based on possible modelling approaches to practical problems which are connected in some way with every-day reality. These studies were used in seventeen experiments with students working in a genuine educational environment under the usual time constraints. Most of the students involved had little or no modelling experience. Results have shown that students have a common set of difficulties, and a set of learning heuristics has been devised in an attempt to overcome these. The theoretical development of formulation-solution processes has identified the following main characteristics in early model development: distribution of features from global (difficult to quantify) to specific (easily quantified) concepts; basic relationships are often generated as solution proceeds; relationships can occur in either general or specific forms; general progress is gauged by relationship 'level'; most variables and constants are generated with relationships; partitioning a problem into sub-problems may be possible initially, but such break-down into distinct parts is often only possible after having seen a pattern of linkages in a relationship level graph. Finally, the implications for assessment methods are examined, and suggestions for further research investigations are made.

Due to their prevalence and associated high rehabilitation costs, this thesis aimed to better understand factors influencing the risk of tibial (TSF) and third metatarsal (MT3SF) stress fractures in Royal Marine recruit training. In Study 1, the standard issue combat assault boot and neutral trainer were assessed during running. Running in the boot caused restricted ankle motion, greater forefoot loading, greater ankle stiffness and a more laterally applied horizontal force vector at the instant of peak braking, suggesting that the risk of incurring MT3SF was greater in this condition. In Study 2, bending stresses were modelled along the length of the third metatarsal of five participants, using individual bone geometry and dynamic gait data. Stresses were modelled for running when barefoot, and when shod in the standard issue footwear. Estimated peak bending stresses were significantly greater in the combat assault boot than the gym trainer, predominantly due to increased plantar loading. Individual bone geometry was however dominant in determining peak bending stresses. In Study 3, a large (n=1065) prospective study was conducted to identify differences in baseline characteristics between recruits sustaining a TSF or MT3SF and those who complete training uninjured. Ten TSF and 14 MT3SF cases were compared to 120 uninjured legs. Results suggest that risk of TSF is greater in those recruits with reduced ability to resist loading and attenuate impact during gait. Results for MT3SF suggest that ankle and foot position at touchdown, and the timing and magnitude of forefoot loading, are important factors influencing risk of this injury. The observation of lower age and BMI in both stress fracture groups was linked to lower bone strength and earlier fatigue mechanisms. This thesis has increased the understanding of MT3SF in particular, and provides information on specific factors which may be associated with MT3SF and TSF in RM recruits during basic training.

Angiogenesis, the process by which new vessels form from existing ones, is a key event in the development of a large and malignant vascularised tumour. A rapid expansion in in vivo and in vitro angiogenesis research in recent years has led to increased knowledge about the processes underlying angiogenesis and to promising steps forward in the development of anti-angiogenic therapies for the treatment of various cancers. However, substantial gaps in knowledge persist and the development of effective treatments remains a major challenge. In this thesis we study tumour-induced angiogenesis within the context of a highly controllable experimental environment: the cornea micropocket assay. Using a multidisciplinary approach that combines experiments, image processing and analysis, and mathematical and computational modelling, we aim to provide mechanistic insight into the action of two angiogenic factors which are known to play central roles during tumour-induced angiogenesis: vascular endothelial growth factor A (VEGF-A) and basic fibroblast growth factor (bFGF). Image analysis techniques are used to extract quantitative data, which are both spatially and temporally resolved, from experimental images. These data are then used to develop and parametrise mathematical models describing the evolution of the corneal vasculature in response to both VEGF-A and bFGF. The first models developed in this thesis are one-dimensional continuum models of angiogenesis in which VEGF-A and/or bFGF are released from a pellet implanted into a mouse cornea. We also use an object-oriented framework, designed to facilitate the re-use and extensibility of hybrid multiscale models of angiogenesis and vascular tumour growth, to develop a complementary three-dimensional hybrid model of the same system. The hybrid model incorporates a new non-local cell sensing model which facilitates the formation of well-perfused and functional vascular networks in three dimensions. The continuum models are used to assess the utility of the cornea micropocket assay as a quantitative assay for angiogenesis, to characterise proposed synergies between VEGF-A and bFGF, and to generate experimentally testable predictions regarding the effect of anti-VEGF-A therapies on bFGF-induced angiogenesis. Meanwhile, the hybrid model is used to provide context for the comparison that is drawn between the continuum models and the data, to study the relative distributions of perfused and unperfused vessels in the evolving neovasculature, and to investigate the impact of tip cell sensing dysregulation on the angiogenic response in the cornea. We have found that by exploiting a close link with quantitative data we have been able to extend the predictive and hypothesis-testing capabilities of our models. As such, this thesis demonstrates the potential for integrating mathematical modelling with image analysis techniques to increase insight into the mechanisms underlying angiogenesis.

A presente pesquisa aborda um problema de programação de veículos rico, em que a característica mais importante é a demanda de múltiplas embarcações para atendimento a uma única tarefa. Trata-se de uma aplicação real do setor de apoio marítimo \"offshore\", das embarcações que fazem o reboque e o lançamento de linhas de ancoragem de sondas de perfuração e unidades de produção. Como método de solução, aplicaram-se duas heurísticas com uma abordagem híbrida que incluem uma inserção baseada em programação linear inteira, visando a minimização do custo total da operação, dentro de um tempo de processamento aceitável. / This research deals with a rich vehicle scheduling problem, having as the most important feature the demand of multiple vessels per task. It is a real problem present in the oil industry related to the vessels that undertake the towing and the launching of mooring lines of drilling and production units. As a solution method, two heuristics with a hybrid approach were applied which include an insertion based on integer linear programming, aiming at minimizing the total cost of the operation, within an acceptable processing time.

Le développement durable, défini à l'origine comme un mode de développement qui répond aux besoins du présent sans compromettre la capacité des générations futures à répondre aux leurs, souffre aujourd'hui des limites de cette définition. On dénombre en effet des centaines de définitions rendant le concept flou au point que mesures politiques et économiques les plus diverses se revendiquent durables.L'objectif de cette thèse est de modéliser le développement durable à l'aide des outils mathématiques de la théorie de la viabilité et de rendre sa définition opérationnelle. La théorie de la viabilité permet en effet d'intégrer les principales caractéristiques d'un développement durable : l'approche multi-objectifs, la prise en compte d'un horizon temporel infini, la possibilité de déterminer plusieurs politiques d'actions viables et le respect de l'équité intergénérationnelle. Le principe de la viabilité est l'étude de systèmes dynamiques contrôlés soumis à un ensemble de contraintes. Le concept central est celui de noyau de viabilité, ensemble regroupant tous les états initiaux à partir desquels il existe une évolution restant dans l'ensemble de contraintes. Après avoir mis en évidence les limites des méthodes utilisées dans la modélisation du développement durable telles que l'analyse multicritères, notre démarche scientifique s'appuie sur le développement de deux modèles mathématiques appliqués à la gestion durable du corridor forestier de Fianarantsoa (Madagascar). L'enjeu est alors de déterminer des politiques d'actions permettant de satisfaire un double objectif : la conservation de la forêt et la satisfaction des besoins de la population locale. Nous démontrons alors l'utilité de la théorie de la viabilité pour atteindre ces objectifs en définissant les contraintes adaptées et en calculant des noyaux de viabilité. / Sustainable development, initially defined as a mode of development that meets present needs without compromising the ability of future generations to meet their own needs, now suffers from the limitations of this definition. Since then, there are hundreds of definitions making the concept so vague that the most diverse political and economic measures claim to be sustainable. The aim of this thesis is to develop a model for sustainable development using the mathematical tools of the viability theory and to make this definition operational. The viability theory enables to integrate the main features of sustainable development : the multi-objective approach, the infinite time horizon, the possibility of determining several viable policies and actions and the respect of generational equity. The principle of viability is the study of dynamical controlled systems subjected to a constraint set. The main concept is the viability kernel, which is the set of all the initial states from which there is an evolution that remain in the constraint set. After underlying the limits of methods used in the modeling of sustainable development such as multi-criteria analysis, our scientific approach is based on the development of two mathematical models applied to the sustainable management of the forest corridor of Fianarantsoa (Madagascar). The challenge is to determine policy actions that meet two objectives : the conservation of the forest and the satisfaction of the needs of local people. Then, we demonstrate the usefulness of the theory of viability to achieve these objectives by defining the appropriate constraints and calculating the viability kernels.

O crescimento de tumores vem chamando a atenção de físicos e matemáticos há mais de sessenta anos. Entretanto, a conversa com biólogos e a interação teoria-experimento têm aparecido apenas recentemente. Equações fenomenológicas e simulações computacionais continuam sendo uma ferramenta comum entre todos os modelos que conhecemos. Assim, nesse trabalho nós estudamos o problema do crescimento de tumores monocamada através das abordagens experimental, teórica e computacional, fortalecendo assim a interação teoria-experimento. Cultivamos células das linhagens HeLa (carcinoma cervical humano), HCT-15 (adenocarcinoma coloretal humano), NIH-HN-13 (carcinoma de células escamosas humanas) e U-251 (glioblastoma neuronal humano), obtendo a dimensão fractal e o comportamento do raio médio com o número de células, além de analisarmos os dados da literatura para a linhagem HT-29 (adenocarcinoma coloretal humano). A seguir nós modelamos a taxa de crescimento do raio médio através de uma curva sigmoidal. A solução analítica dessa equação nos permitiu ajustar bem os dados obtidos experimentalmente, e os parâmetros obtidos serviram para a simulação Monte Carlo dinâmico. Para essa, transformamos a taxa de crescimento do raio em taxa de crescimento do número de células, cujos resultados novamente concordaram muito bem com os dados experimentais. A dimensão fractal dos agregados esteve entre 1; 12 df 1; 21, e concordou com os dados da literatura. Novos resultados foram produzidos: i) O raio médio como uma função do número de células nos permitiu um ajuste do tipo Rc(t) = a[Nc(t) ? N~0]1=2 + R~0, mais geral que a comumente aceita relação Rc(t) = cNc(t)1=2; e ii) os tempos de espera no procedimento MCD se distribuem log-normalmente (ou Gaussianamente em alguns casos), diferentemente da distribuição Poissoniana frequêntemente assumida. A distribuição log-normal nos permitiu também conjecturar que um parâmetro , da relação ht(nT)i / n? T , possa caracterizar o crescimento monocamada de tumores devido à sua estreita abrangência 0; 69 0; 81. Nossos resultados nos permitiram concluir que diferentes condições de cultivo podem gerar diferentes respostas dos parâmetros, além disso, dois fenômenos podem caracterizar esse crescimento no âmbito mesoscópico: A competição por espaços livres e a cooperação entre as células. / Tumor growth has been calling attention of physicists and mathematicians for more than sixty years. However, cross-talking with biologists and the interplay between theory and experiment have emerged just recently. Phenomenological equations and computational simulations are still the common toolbox among all the models we know. Thus, in this work, we have studied the problem of monolayer tumor growth through the experimental, theoretical and computational approaches, enhancing the interaction between theory and experiment. We cultivate HeLa (human cervical carcinoma), HCT-15 (human colorectal adenocarcinoma), NIH-HN-13 (human squamous cell carcinoma) and U-251 (human neuronal glioblastoma) cells, calculating the fractal dimension and the behavior of the mean radius with cell number, and analyzing the literature data from HT-29 (human colorectal adenocarcinoma) lineage. Then we modeled the growth rate of mean radius through a sigmoidal curve. The analytical solution of this equation allowed us to fit well the experimental data and the obtained parameters were used into dynamical Monte Carlo simulation. To do this, we transform the radius growth rate in number of cells growth rate, which again agreed with the experimental data. The fractal dimensions of the aggregates ranged from 1; 12 df 1; 21, and agree with the literature. New findings were produced: i) the mean radius as a function of the number of cells enabled us to adjust the function Rc(t) = a[Nc(t) ? N~0]1=2 + R~0, differently from widely accepted relation Rc(t) = cNc(t)1=2; and ii) the waiting times in the MCD procedure are log-normally distributed (sometimes Gaussian), unlike the Poisson distribution often used. The lognormal distribution also allowed us to conjecture that a parameter , from the power law relation ht(nT)i / n? T , might caracterize the tumor monolayer growth due to its narrow range 0; 69 0; 81. Our findings led us to conclude that different culture conditions may produce different parameter responses, furthermore, two phenomenona can describe the growth in mesoscopic level: the competition for free space and the cooperation between cells.

Durante a solidificação de ligas metálicas, a decantação de grãos equiaxiais pode causar: macrossegregação, mudança na posição da transição colunar-equiaxial e alteração no tamanho de grão equiaxial. O objetivo do presente trabalho é a implementação e análise de um modelo matemático para simular a solidificação sob efeito da decantação de grãos. Este modelo combina: (a) a técnica do autômato celular bi-dimensional em escala microscópica considerando a decantação de grãos, (b) um submodelo microscópico baseado em equações de conservação para simular os processos internos aos grãos e (c) um submodelo macroscópico também baseado em equações de conservação para simular os processos importantes da escala de tamanho macroscópica. Os submodelos macroscópico e microscópico compartilham informações correspondentes às suas escalas características de tamanho através de um método de acoplamento. O modelo matemático completo foi examinado inicialmente através da simulação de diversas situações idealizadas como: a decantação de uma esfera sólida, a decantação de um conjunto de esferas porosas e a decantação e solidificação simultânea de um grão dendrítico isolado e imerso em um líquido super-resfriado. Posteriormente, compararam-se os resultados do modelo para a solidificação isotérmica e para a solidificação unidirecional com resultados disponíveis na literatura obtidos por outros modelos numéricos e por técnicas experimentais. Os resultados do modelo mostraram que a decantação de grãos resulta nos seguintes efeitos no caso da solidificação unidirecional: variação de tamanho médio de grãos e macrossegregação ao longo do lingote e alteração ou até eliminação na posição da transição colunar-equiaxial. / During solidification of metallic alloys, the settling of equiaxed grains can cause macrosegregation, change the position of the columnar-to-equiaxed transition, and change the grain size of equiaxed grains. The objective of the present work is to implement and analyze a mathematical model to simulate the solidification under the effect of grain settling. The present model combines: (a) the cellular automaton technique in two dimensions applied in the microscopic lengthscale to account for the movement of grains; (b) a microscopic submodel based on conservation equations to simulate important phenomena occurring within the grains, and (c) a macroscopic submodel, also based on conservation equations, to simulate important phenomena of the macroscopic lengthscale. Using a strong coupling scheme, the macroscopic and microscopic submodels share important information specific to their typical lengthscales. The complete mathematical model was initially examined in several idealized conditions: the settling of a single solid sphere; the settling of a group of porous spheres; and the growth of a single dendritic grain settling in an undercooled liquid. The model was also used to simulate the isothermal solidification and the unidirectional solidification. For these cases, the outcome of the model was compared with experimental and calculated results available in the literature. For the unidirectional solidification, the model results showed that the settling of grains can cause a variation in the macrosegregation and grain size along the ingot and can change the position or even eliminate the columnar-to-equiaxed transition region.

Neste trabalho foi elaborada uma estrutura de modelos probabilísticos simples que pudessem descrever o processo de tomada de decisão de agentes humanos que são confrontados com a tarefa de prever elementos de uma sequência aleatória gerada por uma cadeia de Markov de memória L. Essa estrutura partiu de uma abordagem bayesiana em que o agente infere uma distribuição de probabilidades a partir de uma série de observações da sequência e de suas próprias respostas, considerando que o agente tenha uma memória de tamanho K. Como resultado da abordagem bayesiana, o agente adota uma estratégia ótima que consiste na perseveração na alternativa mais provável dado o histórico das últimas tentativas; por conta disso e de observações experimentais de que humanos tendem a adotar nesse tipo de problema estratégias sub-ótimas, por exemplo a de pareamento de probabilidades (probability matching), foram desenvolvidas variações sobre esse modelo que tentassem descrever mais de perto o comportamento adotado por humanos. Nesse sentido, foram adotadas as variáveis de troca de resposta (possível ação tomada pelo agente) e de recompensa (possível resultado da ação) na formulação do modelo e foram adicionados parâmetros, inspirados em modelos de ação dopaminérgica, que permitissem um desvio da estratégia ótima resultante da abordagem bayesiana. Os modelos construídos nessa estrutura foram simulados computacionalmente para diversos valores dos parâmetros, incluindo as memórias K e L do agente e da cadeia de Markov, respectivamente. Através de análises de correlação, esses resultados foram comparados aos dados experimentais, de um grupo de pesquisa do Instituto de Ciências Biomédicas da USP, referentes a tarefas de tomada de decisão envolvendo pessoas de diversas faixas etárias (de 3 a 73 anos) e cadeias de Markov de memórias 0, 1 e 2. Nessa comparação, concluiu-se que as diferenças entre grupos etários no experimento podem ser explicadas em nossa modelagem através da variação da memória K do agente crianças de até 5 anos mostram um limite K = 1, e as de até 12 anos mostram um limite K = 2 e da variação de um parâmetro de reforço de aprendizado dependendo do grupo e da situação de decisão à qual os indivíduos eram expostos, o valor ajustado desse parâmetro variou de 10% para baixo até 30% para cima do seu valor original de acordo com a abordagem bayesiana. / In this work we developed a simple probabilistic modeling framework that could describe the process of decision making in human agents that are presented with the task of predicting elements of a random sequence generated by a Markov chain with memory L. Such framework arised from a Bayesian approach in which the agent infers a probability distribution from a series of observations on the sequence and on its own answers, and considers that the agent\'s memory has length K. As a result of the Bayesian approach, the agent adopts an optimal strategy that consists in perseveration of the most likely alternative given the history of the last few trials; because of that and of experimental evidence that humans tend, in such kinds of problems, to adopt suboptimal strategies such as probability matching, variations on that model were developed in an attempt to have a closer description of the behavior adopted by humans. In that sense, the `shift\' (possible action taken by the agent on its response) and `reward\' (possible result of the action) variables were adopted in the formulation of the model, and parameters inspired by models of dopaminergic action were added to allow deviation from the optimal strategy that resulted from the Bayesian approach. The models developed in that framework were computationally simulated for many values of the parameters, including the agent\'s and the Markov chain\'s memory lengths K and L respectively. Through correlation analysis these results were compared to experimental data, from a research group from the Biomedical Science Institute at USP, regarding decision making tasks that involved people of various ages (3 to 73 years old) and Markov chains of orders 0, 1 and 2. In this comparison it was concluded that the differences between age groups in the experiment can be explained in our modeling through variation of the agent\'s memory length K children up to 5 years old exhibited a limitation of K = 1, and those up to 12 years old were limited to K = 2 and through variation of a learning reinforcement parameter depending on the group and the decision situation to which the candidates were exposed, the fitted value for that parameter ranged from 10% below to 30% above its original value according to the Bayesian approach.

Cadeias de suprimentos podem ter operações seguindo diferentes estratégias de distribuição e a utilização de cada uma dessas estratégias pode resultar em diferentes operações e custos. A estratégia de cross-docking auxilia na redução dos custos de distribuição de produtos, consolidando cargas, e a redução de tempo e custos de armazenamento, uma vez que o tempo máximo de estoque permitido pela estratégia é de cerca de 24 horas. O objetivo deste trabalho é apresentar um modelo para o problema de cross-docking, em que cargas são entregues e reorganizadas de forma a atender a outras cargas que são coletadas e garantir que as janelas de tempo para início das operações sejam atendidas. Devido à falta de instâncias para o problema disponíveis na literatura, buscou-se gerar um benchmark e disponibilizá-las à comunidade científica. Uma vez que o problema é de difícil solução exata, um método heurístico para a resolução do problema foi desenvolvido. Os resultados mostraram que o modelo proposto resulta em boas soluções quando comparado ao modelo da literatura. O estudo de calibração do software IBM CPLEX mostrou que a calibração dos parâmetros pode resultar em melhores soluções e, por fim, a matheurística se mostrou competitiva com o CPLEX, principalmente para cenários em que a proporção de entregas e coletas diverge. / Supply chains may have operations which follow different distribution strategies and each one of these strategies may result in different operations and costs. The Cross-docking strategy helps to reduce the products distribution costs by consolidating loads and reducing storage costs as the maximum inventory time is approximately 24 hours. The aim of this research is to present a model for the cross-docking problem where loads are delivered and reorganized so as to cater for other loads that are collected and ensure that time windows are respected. Due to the lack of instances available in the literature, a benchmark was generated and was made available to the scientific community. As the problem is difficult to obtain the exact solution, a heuristic method was developed. The results showed that the proposed model has good solutions when compared to the literature model. A study of the IBM CPLEX software showed that tuning can result in better solutions and the matheuristcs was competitive with the software, mainly in scenarios where deliveries and pickups are very different.

Esta dissertação trata da hidráulica da queda livre em canal de seção retangular. A análise bibliográfica do tema está calcada nos trabalhos pioneiros, nos clássicos e nos recentes. Com base nos princípios da Física: Conservação de Massa, Quantidade de Movimento, e Primeira Lei da Termodinâmica, foi desenvolvido um modelo matemático para a queda livre. O modelo proposto, após ser analisado do ponto de vista de sua consistência, foi validado em comparações com resultados fornecidos por outros pesquisadores, geralmente com modelos empíricos (ou semi-empíricos), ajustados a dados experimentais. Os resultados obtidos pelo modelo matemático proposto nesta dissertação, correspondem a boas estimativas das grandezas envolvidas nos escoamentos em queda livre, o que credencia o modelo proposto como uma ferramenta apropriada para projetos em engenharia hidráulica, principalmente quando se tem em conta que a queda livre é o mecanismo de dissipação mais presente na natureza. / This study addresses free fall hydraulics in rectangular channel section. References were based on not only earlier studies, but also on classical and most recent ones. Based on principles of Physics, such as mass conservation, momentum and the First Law of Thermodynamics, a mathematical model has been developed as an example of free fall hydraulic. After extensive consistency analyses the proposed model has been validated by comparing different results furnished by other researchers, generally based on empirical or semiempirical treatment adjusted to experimental data. Results obtained from the mathematical model proposed here correspond to good estimates of greatnesses involved in the free overfall and that turns the proposed model into an adequate tool for Hydraulic Engineering projects, especially when we all know free overfall is the most common dissipation mechanism in Nature.