Um conto, um canto, um encanto: a história, o mapa, a música em Belo Horizonte

O presente trabalho pretende ser uma contribuição no campo da etnomusicologia, e resulta de um estudo e pesquisa de campo realizada na cidade de Belo Horizonte, capital do estado de Minas Gerais, Brasil. Belo Horizonte é uma cidade “inventada”, assim como Brasília, a capital federal. Enquanto cidade, Belo Horizonte foi desenhada, projetada, nasceu de uma prancheta e foi oficialmente inaugurada em 12 de dezembro de 1897. Nesta tese tem como objetivo compreender e demonstrar como as transformações sociais e espaciais foram provocadas pelo acontecimento musical. Defendo a posição que o entendimento do que denomino Acontecimento Musical permite perceber o potencial da música de reorganizar normas, inicialmente defendidas pelos idealizadores da cidade, da utilização e ocupação dos lugares e espaços, e construindo, pela prática musical, uma nova cartografia da cidade, que se impõe a cartografia oficialmente idealizada, resultando em uma Cartografia Musical de Belo Horizonte.

Stratégie de gestion de stocks dans un réseau collaboratif

Le contenu de ce mémoire traite du problème de gestion des stocks dans un réseau constitué de plusieurs sites de stockage de produits. Chaque site i gère son stock de manière autonome pour satisfaire une demande déterministe sur un horizon donné. Un stock maximum Si est tenu à chaque site i. Lorsque le point de commande si est atteint, une commande de taille Qi est placée au centre de distribution qui alimente tous les sites. Qi est telle que Qi = Si - si. La quantité Qi est livrée dans un délai connu Li. Si, à un instant donné, la demande Di au site i excède la quantité en main, le site i fait appel à un ou à plusieurs autres sites du réseau pour le transfert d’une quantité Xji (j = 1, 2, …, n). Ce transfert s’effectue selon un certain nombre de règles de jeu qui tiennent compte des coûts de transfert, de stockage, d’approvisionnement et de pénurie. Ce mémoire examine six principales contributions publiées dans la littérature pour évaluer les contributions d’un modèle collaboratif aux performances, en termes de coûts et de niveau de service, de chaque site du réseau. Cette investigation se limite à une configuration du réseau à deux échelons : un entrepôt central et n (n > 2) sites de stockage. Le cas des pièces de rechange, caractérisé par une demande aléatoire, est examiné dans trois chapitres de ce mémoire. Une autre application de ces stratégies à la collaboration entre n centres hospitaliers (n > 2) est également examinée dans ce travail.

Decision making models embedded into a web-based tool for assessing pest infestation risk

Current practices in agricultural management involve the application of rules and techniques to ensure high quality and environmentally friendly production. Based on their experience, agricultural technicians and farmers make critical decisions affecting crop growth while considering several interwoven agricultural, technological, environmental, legal and economic factors. In this context, decision support systems and the knowledge models that support them, enable the incorporation of valuable experience into software systems providing support to agricultural technicians to make rapid and effective decisions for efficient crop growth. Pest control is an important issue in agricultural management due to crop yield reductions caused by pests and it involves expert knowledge. This paper presents a formalisation of the pest control problem and the workflow followed by agricultural technicians and farmers in integrated pest management, the crop production strategy that combines different practices for growing healthy crops whilst minimising pesticide use. A generic decision schema for estimating infestation risk of a given pest on a given crop is defined and it acts as a metamodel for the maintenance and extension of the knowledge embedded in a pest management decision support system which is also presented. This software tool has been implemented by integrating a rule-based tool into web-based architecture. Evaluation from validity and usability perspectives concluded that both agricultural technicians and farmers considered it a useful tool in pest control, particularly for training new technicians and inexperienced farmers.

Big data: appearance and disappearance

The mainstay of Big Data is prediction in that it allows practitioners, researchers, and policy analysts to predict trends based upon the analysis of large and varied sources of data. These can range from changing social and political opinions, patterns in crimes, and consumer behaviour. Big Data has therefore shifted the criterion of success in science from causal explanations to predictive modelling and simulation. The 19th-century science sought to capture phenomena and seek to show the appearance of it through causal mechanisms while 20th-century science attempted to save the appearance and relinquish causal explanations. Now 21st-century science in the form of Big Data is concerned with the prediction of appearances and nothing more. However, this pulls social science back in the direction of a more rule- or law-governed reality model of science and away from a consideration of the internal nature of rules in relation to various practices. In effect Big Data offers us no more than a world of surface appearance and in doing so it makes disappear any context-specific conceptual sensitivity.

”Stress Monitoring via Wearable Electrodermal Sensors”

Stress is a phenomenon that on some level affects everyone’s lives on a daily basis. The autonomic nervous system controls the varying levels of stress at any given time. The responses of the autonomic nervous system adjust the body to cope with changing external and internal conditions. During high-stress situations the body is forced into a state of heightened alertness, which passes when the stressor is removed. The stressor can be any external or internal event that causes the body to respond. Stress is a very versatile phenomenon that can be both a cause and an indicator of other medical conditions, for example cardiovascular disease. Stress detection can therefore be helpful in identifying these conditions and monitoring the overall emotional state of a person. Electrodermal activity (EDA) is one of the most easily implemented ways to monitor the activity of the autonomic nervous system. EDA describes changes occurring in the various electrical properties of the skin, including skin conductivity and resistance. Increased emotional sweating has been proven to be one possible indication of stress. On the surface of the skin, increased sweating translates to increased skin conductivity, which can be observed through EDA measurements. This makes electrodermal activity a very useful tool in a wide range of applications where it is desirable to observe changes in a person’s stress level. EDA can be recorded by using specialized body sensors placed on specific locations on the body. Most commonly used recording sites are the palms of the hands due to the high sweat gland density on those areas. Measurement is done using at least two electrodes attached to the skin, and recording the electrical conductance between them. This thesis implements a prototype of a wireless EDA measurement system. The feasibility of the prototype is also verified with a small group of test subjects. EDA was recorded from the subjects while they were playing a game of Tetris. The goal was to observe variations in the measured EDA that would indicate changes in the subjects’ stress levels during the game. The analysis of the obtained measurement results confirmed the connection between stress and recorded EDA. During the game, random occurrences of lowered skin resistance were clearly observable, which indicates points in the game where the player felt more anxious. A wireless measurement system has the potential of offering more flexible and comfortable long-term measuring of EDA, and could be utilized in a wide range of applications.

Uma abordagem evolucionária e espacial para o jogo da autorregulação de processos de trocas sociais em sistemas multiagentes

Interações sociais são frequentemente descritas como trocas sociais. Na literatura, trocas sociais em Sistemas Multiagentes são objeto de estudo em diversos contextos, nos quais as relações sociais são interpretadas como trocas sociais. Dentre os problemas estudados, um problema fundamental discutido na literatura e a regulação¸ ao de trocas sociais, por exemplo, a emergência de trocas equilibradas ao longo do tempo levando ao equilíbrio social e/ou comportamento de equilíbrio/justiça. Em particular, o problema da regulação de trocas sociais e difícil quando os agentes tem informação incompleta sobre as estratégias de troca dos outros agentes, especificamente se os agentes tem diferentes estratégias de troca. Esta dissertação de mestrado propõe uma abordagem para a autorregulacao de trocas sociais em sistemas multiagentes, baseada na Teoria dos Jogos. Propõe o modelo de Jogo de Autorregulacão ao de Processos de Trocas Sociais (JAPTS), em uma versão evolutiva e espacial, onde os agentes organizados em uma rede complexa, podem evoluir suas diferentes estratégias de troca social. As estratégias de troca são definidas através dos parâmetros de uma função de fitness. Analisa-se a possibilidade do surgimento do comportamento de equilíbrio quando os agentes, tentando maximizar sua adaptação através da função de fitness, procuram aumentar o numero de interações bem sucedidas. Considera-se um jogo de informação incompleta, uma vez que os agentes não tem informações sobre as estratégias de outros agentes. Para o processo de aprendizado de estratégias, utiliza-se um algoritmo evolutivo, no qual os agentes visando maximizar a sua função de fitness, atuam como autorregulares dos processos de trocas possibilitadas pelo jogo, contribuindo para o aumento do numero de interações bem sucedidas. São analisados 5 diferentes casos de composição da sociedade. Para alguns casos, analisa-se também um segundo tipo de cenário, onde a topologia de rede é modificada, representando algum tipo de mobilidade, a fim de analisar se os resultados são dependentes da vizinhança. Alem disso, um terceiro cenário é estudado, no qual é se determinada uma política de influencia, quando as medias dos parâmetros que definem as estratégias adotadas pelos agentes tornam-se publicas em alguns momentos da simulação, e os agentes que adotam a mesma estratégia de troca, influenciados por isso, imitam esses valores. O modelo foi implementado em NetLogo.

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

'Bayesian Optimisation for Nurse Scheduling'

A Bayesian optimisation algorithm for a nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. When a human scheduler works, he normally builds a schedule systematically following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not yet completed, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this paper, we design a more human-like scheduling algorithm, by using a Bayesian optimisation algorithm to implement explicit learning from past solutions. A nurse scheduling problem from a UK hospital is used for testing. Unlike our previous work that used Genetic Algorithms to implement implicit learning [1], the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The Bayesian optimisation algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, new rule strings have been obtained. Sets of rule strings are generated in this way, some of which will replace previous strings based on fitness. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. For clarity, consider the following toy example of scheduling five nurses with two rules (1: random allocation, 2: allocate nurse to low-cost shifts). In the beginning of the search, the probabilities of choosing rule 1 or 2 for each nurse is equal, i.e. 50%. After a few iterations, due to the selection pressure and reinforcement learning, we experience two solution pathways: Because pure low-cost or random allocation produces low quality solutions, either rule 1 is used for the first 2-3 nurses and rule 2 on remainder or vice versa. In essence, Bayesian network learns 'use rule 2 after 2-3x using rule 1' or vice versa. It should be noted that for our and most other scheduling problems, the structure of the network model is known and all variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus, learning can amount to 'counting' in the case of multinomial distributions. For our problem, we use our rules: Random, Cheapest Cost, Best Cover and Balance of Cost and Cover. In more detail, the steps of our Bayesian optimisation algorithm for nurse scheduling are: 1. Set t = 0, and generate an initial population P(0) at random; 2. Use roulette-wheel selection to choose a set of promising rule strings S(t) from P(t); 3. Compute conditional probabilities of each node according to this set of promising solutions; 4. Assign each nurse using roulette-wheel selection based on the rules' conditional probabilities. A set of new rule strings O(t) will be generated in this way; 5. Create a new population P(t+1) by replacing some rule strings from P(t) with O(t), and set t = t+1; 6. If the termination conditions are not met (we use 2000 generations), go to step 2. Computational results from 52 real data instances demonstrate the success of this approach. They also suggest that the learning mechanism in the proposed approach might be suitable for other scheduling problems. Another direction for further research is to see if there is a good constructing sequence for individual data instances, given a fixed nurse scheduling order. If so, the good patterns could be recognized and then extracted as new domain knowledge. Thus, by using this extracted knowledge, we can assign specific rules to the corresponding nurses beforehand, and only schedule the remaining nurses with all available rules, making it possible to reduce the solution space. Acknowledgements The work was funded by the UK Government's major funding agency, Engineering and Physical Sciences Research Council (EPSRC), under grand GR/R92899/01. References [1] Aickelin U, "An Indirect Genetic Algorithm for Set Covering Problems", Journal of the Operational Research Society, 53(10): 1118-1126,

As habilidades sociais em dependentes de álcool: Um estudo exploratório

A vida social de um indivíduo dependente do álcool é na maioria das vezes um factor de risco para continuar ou aumentar o consumo excessivo de bebidas alcoólicas. Um dos grandes fracassos do alcoólico é não cumprir adequadamente um papel social desejado, o que resulta em prejuízos para si mesmo e para os outros. O indivíduo que abusa no consumo, depressa perde a sua reputação junto de colegas, amigos e familiares, o que o deixa mais intolerante à frustração e aumenta o consumo. A mentira toma-se então sua aliada, pois através dela ele vai reduzindo a ansiedade causada pelo fracasso na vida social e que os outros teimam em deixar bem nítido. Identificar os problemas sociais dos quais o indivíduo padece, é fundamental para planear melhor uma estratégia de intervenção, quer seja ela de prevenção, de psicoterapia ou de reabilitação. Os programas de tratamento habitualmente propostos para a abordagem dos problemas derivados do consumo de álcool centram a sua atenção, quase exclusivamente, no comportamento aditivo como guia orientador da intervenção e como indicador objectivo do êxito do próprio programa Mas na maioria dos casos o comportamento aditivo é sim a manifestação mais objectiva de um profundo desajuste entre o sujeito consigo mesmo e com o seu meio ambiente. É por isso, objectivo dos processos de recuperação oferecer-lhe a possibilidade de recuperar a crença na palavra ou aprender o seu valor como meio de comunicação fundamental entre os homens. Para além de possibilitar aos sujeitos dependentes de álcool este valor, importa também incutir nos sujeitos o valor positivo de viver com limites; pois são especialistas em tentar sabotar a acção dos técnicos e em descobrir as suas debilidades para as utilizarem em seu interesse. Importa por isso, que aprendam o valor das leis e a utilidade, para todos, de cumpri-las (Kalina, 2001). Assim, o treino de habilidades sociais constitui uma parte importante dos tratamentos para os sujeitos com problemas de abuso de álcool e drogas. Foi nesse sentido que nos propusemos a identificar o nível de habilidades sociais em pessoas dependentes de álcool. O estudo que desenvolvemos é de carácter exploratório/descritivo, para o qual optámos por utilizar uma metodologia quantitativa. A amostra foi constituída por 229 indivíduos, do sexo masculino, dependentes de álcool, em instituições nacionais de referência na área da alcoologia O instrumento de recolha de dados é constituído por um Questionário de dados sócio demográficos, uma Escala de Habilidades Sociais e uma Escala de Auto-apreciação Pessoal. Constatamos que a amostra constituída por indivíduos dependentes de álcool apresenta uma pontuação média na Escala de Habilidades Sociais de 89.96, equivalente ao percentil 55 na tabela de parametrização de Gismero (2002). Este valor é claramente inferior ao conseguido por qualquer uma das outras amostras analisadas, seja a do estudo preliminar, seja a do estudo comparativo, constituída por indivíduos da população em geral e que conseguiram um percentil 70. ABSTRACT; The social life of a person dependent on alcohol is, most of the time, a risk factor to continue or increase the alcohol excessive consumption. One of the alcoholic failures is the fact that he is unable to perform an adequate social role, to the detriment of himself and others. A person, who abuses alcohol consumption, soon loses his reputation next to his colleagues, friends and relatives, which makes him intolerant of frustration and increases the alcohol consumption. To lie becomes his best ally, because it helps him to reduce the anxiety caused by the failure of his social life, what is promptly pointed out by others. To identify the individual social problems is essential to plan the best intervention strategy. This can be of prevention, psychotherapy or rehabilitation. The treatment programmers, usually proposed to deal with the problems caused by alcohol consumption, focus almost exclusively on the addictive behaviour, as a guide line for the intervention and as an objective indicator of the success of the programme itself. But, in most cases, the addictive behaviour is an objective manifestation of a deep break off of the individual with himself and with his environment. That is why the aim of the recuperation process is to offer the individual the possibility to recover their belief on the word or to learn its value as an essential means of communication for men. Besides getting the message trough, it is also important to make the individuals aware of the positive value of living within limits. These individuals are specialists on trying to sabotage the technicians’ actions, discovering their weaknesses so they can use them on their own behalf. That is why it is so important that they learn the value of rules and the importance of accomplishing them (Kalina, 2001). Therefore, the training of the social skills is an important part of the treatment of individuals with problems of alcohol or dugs addiction. So, we committed ourselves to identifying the level of social skills on people who have an alcohol addiction. The study we developed is exploratory/ descriptive and we chose to use a quantitative methodology. The sample was of 229 male alcohol dependent individuals, staying in national institutes of reference in the area of alcohol abuse and alcoholism. The means to collect data were a social demographic data questionnaire, a scale of social skills and a scale of personal self- assessment. We realized that the sample of alcohol dependent individuals presents an average score in social skills of 89.96, equivalent to a percentile of 55 in the parameterization of Gismero (2002). This is clearly a lower value than the one obtained by any other sample we analyzed, whether in the preliminary study or in the comparative study, constituted by individuals of the common population that achieved a percentile of 70.

Estrutura e processo de formação das formas culturais: o caso do cante alentejano

Qualquer sistema de práticas individualizado que se distinga aos olhos dos seus praticantes é caracterizável pela forma que assume e pelo seu conjunto de atributos. A essa forma corresponde uma moldura analítica que encontra o seu equivalente real nos actores e no modo como formulam discursos sobre as suas práticas, do qual emerge um conjunto de normas. Por conseguinte, a forma tanto é uma construção metodológica externa para a descrição do sistema e das suas condições de existência como uma realidade dinâmica, produtora de identidades culturais. Pretendemos substituir uma noção imprecisa de “forma cultural” por um conceito estruturado, definindo-a como sistema de referência que os membros de uma cultura partilham e que define e regula as produções e reproduções culturais e que comporta um elemento estruturante, um sistema normativo e uma dinâmica social. O caminho para essa conceptualização implica a aplicação do conceito a um objecto empírico, operação que realizamos ao analisar o Cante Alentejano – conjunto de maneiras de cantar observadas do Alentejo – enquanto forma cultural; Abstract: Any system of practices that can be individualized and distinguished by its practitioners can be characterized through the form it assumes and the set of its attributes. Such form corresponds to an analytical framework that has its equivalent in the real actors world and in the ways they formulate utterances about their practices, from which emerges a set of rules. Therefore, the form is both an external methodological construction needed for the system’s description and a dynamic “emic” reality that produces cultural identities. We intend to replace an inaccurate notion of "cultural form" by a structured concept. We will define it as the reference system that the members of a given culture share and that guides and regulates the cultural processes of production and reproductions of the system. The concept comprises a structural element, a normative system, and a social dynamic. The path to this conceptualization implies applying the concept to an empirical object, operation that will be held by analyzing Cante Alentejano – a set of ways of singing from Alentejo – as a cultural form.

Bayesian Optimisation Algorithm for Nurse Scheduling, Scalable Optimization via Probabilistic Modeling

Our research has shown that schedules can be built mimicking a human scheduler by using a set of rules that involve domain knowledge. This chapter presents a Bayesian Optimization Algorithm (BOA) for the nurse scheduling problem that chooses such suitable scheduling rules from a set for each nurse’s assignment. Based on the idea of using probabilistic models, the BOA builds a Bayesian network for the set of promising solutions and samples these networks to generate new candidate solutions. Computational results from 52 real data instances demonstrate the success of this approach. It is also suggested that the learning mechanism in the proposed algorithm may be suitable for other scheduling problems.

Práticas de ensino supervisionada e a importância da creche na transição para o Jardim-de-Infância

O presente Relatório Final de Estágio foi desenvolvido no âmbito da unidade curricular de Prática de Ensino Supervisionada III (PES III) do 2.º Ciclo de Estudos, do Mestrado em Educação Pré-Escolar e Ensino do 1.º Ciclo do Ensino Básico e inclui duas partes distintas. A primeira parte diz respeito às práticas de estágio desenvolvidas no decorrer da unidade curricular Prática de Ensino Supervisionada (PES). Aí, proceder-se-á a uma reflexão crítica acerca do trabalho desenvolvido nos referidos contextos e das experiências de aprendizagem que daí resultaram. A segunda parte do relatório final de estágio está relacionada com a execução de um trabalho de investigação centrado nos seguintes objetivos: compreender a importância da frequência da creche no desenvolvimento e aprendizagem das crianças, identificar a opinião dos profissionais sobre a creche e as suas atividades e conhecer a relação pais/Jardim-de-Infância/Creche. O estudo em causa segue uma metodologia qualitativa, mais concretamente o estudo de caso e envolve duas educadoras de creche, três educadoras do jardim-de-infância, nove pais das crianças da creche e sessenta e uma crianças (9 responderam às entrevistas acerca da frequência da creche e 52 executaram o desenho alusivo à visita à sala de creche). Com a realização deste trabalho de investigação pretende-se acima de tudo obter resposta à seguinte questão: “Qual a importância da frequência da Creche na transição para o Jardim-de-Infância?” Tendo em conta o estudo realizado foi possível obter determinadas conclusões de acordo com as perpetivas dos profissionais (educadoras de creche e educadoras do jardim-de-infância), pais e crianças do jardim-de-infância. No parecer dos profissionais a creche é importante no sentido de autonomia (gerir a separação dos pais), em termos de regras, altruísmo e partilha de brinquedos, bem como o desenvolvimento global das crianças. Os pais consideram que a creche proporciona o desenvolvimento pessoal e aprendizagem das crianças. Para além disto, a mesma permite educar os seus filhos e é a primeira etapa da educação escolar. Porém, existem pais que referem que o facto de deixarem os seus filhos na referida instituição está relacionado com a inexistência de tempo da sua parte.De acordo com as entrevistas realizadas às crianças do jardim-de-infância podemos verificar que a frequência da creche foi fundamental para o crescimento saudável das crianças, sendo que 89% das crianças têm lembranças positivas sobre o primeiro dia de creche, 57% das crianças menciona como atividades que mais gostavam brincar com as bolas, 100% mencionaram que preferem brincar com outras crianças, 78% revelaram a existência de uma interação positiva entre as mesmas e a educadora, bem como 100% afirmou existir reações positivas à participação dos pais nas suas atividades.

'Bayesian Optimisation for Nurse Scheduling'

A Bayesian optimisation algorithm for a nurse scheduling problem is presented, which involves choosing a suitable scheduling rule from a set for each nurse's assignment. When a human scheduler works, he normally builds a schedule systematically following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not yet completed, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this paper, we design a more human-like scheduling algorithm, by using a Bayesian optimisation algorithm to implement explicit learning from past solutions. A nurse scheduling problem from a UK hospital is used for testing. Unlike our previous work that used Genetic Algorithms to implement implicit learning [1], the learning in the proposed algorithm is explicit, i.e. we identify and mix building blocks directly. The Bayesian optimisation algorithm is applied to implement such explicit learning by building a Bayesian network of the joint distribution of solutions. The conditional probability of each variable in the network is computed according to an initial set of promising solutions. Subsequently, each new instance for each variable is generated by using the corresponding conditional probabilities, until all variables have been generated, i.e. in our case, new rule strings have been obtained. Sets of rule strings are generated in this way, some of which will replace previous strings based on fitness. If stopping conditions are not met, the conditional probabilities for all nodes in the Bayesian network are updated again using the current set of promising rule strings. For clarity, consider the following toy example of scheduling five nurses with two rules (1: random allocation, 2: allocate nurse to low-cost shifts). In the beginning of the search, the probabilities of choosing rule 1 or 2 for each nurse is equal, i.e. 50%. After a few iterations, due to the selection pressure and reinforcement learning, we experience two solution pathways: Because pure low-cost or random allocation produces low quality solutions, either rule 1 is used for the first 2-3 nurses and rule 2 on remainder or vice versa. In essence, Bayesian network learns 'use rule 2 after 2-3x using rule 1' or vice versa. It should be noted that for our and most other scheduling problems, the structure of the network model is known and all variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus, learning can amount to 'counting' in the case of multinomial distributions. For our problem, we use our rules: Random, Cheapest Cost, Best Cover and Balance of Cost and Cover. In more detail, the steps of our Bayesian optimisation algorithm for nurse scheduling are: 1. Set t = 0, and generate an initial population P(0) at random; 2. Use roulette-wheel selection to choose a set of promising rule strings S(t) from P(t); 3. Compute conditional probabilities of each node according to this set of promising solutions; 4. Assign each nurse using roulette-wheel selection based on the rules' conditional probabilities. A set of new rule strings O(t) will be generated in this way; 5. Create a new population P(t+1) by replacing some rule strings from P(t) with O(t), and set t = t+1; 6. If the termination conditions are not met (we use 2000 generations), go to step 2. Computational results from 52 real data instances demonstrate the success of this approach. They also suggest that the learning mechanism in the proposed approach might be suitable for other scheduling problems. Another direction for further research is to see if there is a good constructing sequence for individual data instances, given a fixed nurse scheduling order. If so, the good patterns could be recognized and then extracted as new domain knowledge. Thus, by using this extracted knowledge, we can assign specific rules to the corresponding nurses beforehand, and only schedule the remaining nurses with all available rules, making it possible to reduce the solution space. Acknowledgements The work was funded by the UK Government's major funding agency, Engineering and Physical Sciences Research Council (EPSRC), under grand GR/R92899/01. References [1] Aickelin U, "An Indirect Genetic Algorithm for Set Covering Problems", Journal of the Operational Research Society, 53(10): 1118-1126,

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

Introduction to Multi-Agent Simulation

When designing systems that are complex, dynamic and stochastic in nature, simulation is generally recognised as one of the best design support technologies, and a valuable aid in the strategic and tactical decision making process. A simulation model consists of a set of rules that define how a system changes over time, given its current state. Unlike analytical models, a simulation model is not solved but is run and the changes of system states can be observed at any point in time. This provides an insight into system dynamics rather than just predicting the output of a system based on specific inputs. Simulation is not a decision making tool but a decision support tool, allowing better informed decisions to be made. Due to the complexity of the real world, a simulation model can only be an approximation of the target system. The essence of the art of simulation modelling is abstraction and simplification. Only those characteristics that are important for the study and analysis of the target system should be included in the simulation model. The purpose of simulation is either to better understand the operation of a target system, or to make predictions about a target system’s performance. It can be viewed as an artificial white-room which allows one to gain insight but also to test new theories and practices without disrupting the daily routine of the focal organisation. What you can expect to gain from a simulation study is very well summarised by FIRMA (2000). His idea is that if the theory that has been framed about the target system holds, and if this theory has been adequately translated into a computer model this would allow you to answer some of the following questions: · Which kind of behaviour can be expected under arbitrarily given parameter combinations and initial conditions? · Which kind of behaviour will a given target system display in the future? · Which state will the target system reach in the future? The required accuracy of the simulation model very much depends on the type of question one is trying to answer. In order to be able to respond to the first question the simulation model needs to be an explanatory model. This requires less data accuracy. In comparison, the simulation model required to answer the latter two questions has to be predictive in nature and therefore needs highly accurate input data to achieve credible outputs. These predictions involve showing trends, rather than giving precise and absolute predictions of the target system performance. The numerical results of a simulation experiment on their own are most often not very useful and need to be rigorously analysed with statistical methods. These results then need to be considered in the context of the real system and interpreted in a qualitative way to make meaningful recommendations or compile best practice guidelines. One needs a good working knowledge about the behaviour of the real system to be able to fully exploit the understanding gained from simulation experiments. The goal of this chapter is to brace the newcomer to the topic of what we think is a valuable asset to the toolset of analysts and decision makers. We will give you a summary of information we have gathered from the literature and of the experiences that we have made first hand during the last five years, whilst obtaining a better understanding of this exciting technology. We hope that this will help you to avoid some pitfalls that we have unwittingly encountered. Section 2 is an introduction to the different types of simulation used in Operational Research and Management Science with a clear focus on agent-based simulation. In Section 3 we outline the theoretical background of multi-agent systems and their elements to prepare you for Section 4 where we discuss how to develop a multi-agent simulation model. Section 5 outlines a simple example of a multi-agent system. Section 6 provides a collection of resources for further studies and finally in Section 7 we will conclude the chapter with a short summary.