'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.

Sensory gating and its modulation by cannabinoids: electrophysiological, computational and mathematical analysis

Gating of sensory information can be assessed using an auditory conditioning-test paradigm which measures the reduction in the auditory evoked response to a test stimulus following an initial conditioning stimulus. Recording brainwaves from specific areas of the brain using multiple electrodes is helpful in the study of the neurobiology of sensory gating. In this paper, we use such technology to investigate the role of cannabinoids in sensory gating in the CA3 region of the rat hippocampus. Our experimental results show that application of the exogenous cannabinoid agonist WIN55,212-2 can abolish sensory gating. We have developed a phenomenological model of cannabinoid dynamics incorporated within a spiking neural network model of CA3 with synaptically interacting pyramidal and basket cells. Direct numerical simulations of this model suggest that the basic mechanism for this effect can be traced to the suppression of inhibition of slow GABAB synapses. Furthermore, by working with a simpler mathematical firing rate model we are able to show the robustness of this mechanism for the abolition of sensory gating.

Conceptual guidelines for the international expansion of SMEs from Curaçao

Internationalization and moving to new markets can create many opportunities for small businesses across the globe, but also presents a number of new challenges they will face, which may influence their competitive advantage in the global market -- Present paper aims to provide an internationalization guide for SMEs from Curaçao -- Also the determinants that can impact internationalization will be discussed -- In this paper, three widely researched internationalization models form the basis of the theoretical perspectives of this paper; the traditional Uppsala model, the Network model and the Linkage, Leverage and Learning model

Fundamentals for the Development of a Performance Analysis Approach in Collaborative Networks

Part 15: Performance Management Frameworks

Detecção do mal-posicionamento rotacional de dedos em dispositivos de captura de impressões digitais multivista sem toque utilizando redes neurais artificiais

Dissertação (mestrado)—Universidade de Brasília, Faculdade de Tecnologia, Departamento de Engenharia Mecânica, 2016.

Controle preditivo neural aplicado à processos petroquímicos

A pesquisa tem como objetivo desenvolver uma estrutura de controle preditivo neural, com o intuito de controlar um processo de pH, caracterizado por ser um sistema SISO (Single Input - Single Output). O controle de pH é um processo de grande importância na indústria petroquímica, onde se deseja manter constante o nível de acidez de um produto ou neutralizar o afluente de uma planta de tratamento de fluidos. O processo de controle de pH exige robustez do sistema de controle, pois este processo pode ter ganho estático e dinâmica nãolineares. O controlador preditivo neural envolve duas outras teorias para o seu desenvolvimento, a primeira referente ao controle preditivo e a outra a redes neurais artificiais (RNA s). Este controlador pode ser dividido em dois blocos, um responsável pela identificação e outro pelo o cálculo do sinal de controle. Para realizar a identificação neural é utilizada uma RNA com arquitetura feedforward multicamadas com aprendizagem baseada na metodologia da Propagação Retroativa do Erro (Error Back Propagation). A partir de dados de entrada e saída da planta é iniciado o treinamento offline da rede. Dessa forma, os pesos sinápticos são ajustados e a rede está apta para representar o sistema com a máxima precisão possível. O modelo neural gerado é usado para predizer as saídas futuras do sistema, com isso o otimizador calcula uma série de ações de controle, através da minimização de uma função objetivo quadrática, fazendo com que a saída do processo siga um sinal de referência desejado. Foram desenvolvidos dois aplicativos, ambos na plataforma Builder C++, o primeiro realiza a identificação, via redes neurais e o segundo é responsável pelo controle do processo. As ferramentas aqui implementadas e aplicadas são genéricas, ambas permitem a aplicação da estrutura de controle a qualquer novo processo

Um sistema para a inferência de variáveis usando redes neurais artificiais

This work aims to obtain a low-cost virtual sensor to estimate the quality of LPG. For the acquisition of data from a distillation tower, software HYSYS ® was used to simulate chemical processes. These data will be used for training and validation of an Artificial Neural Network (ANN). This network will aim to estimate from available simulated variables such as temperature, pressure and discharge flow of a distillation tower, the mole fraction of pentane present in LPG. Thus, allowing a better control of product quality

WATER USE AND SYSTEM RELIABILITY UNDER DIESELGENERATOR AND SOLAR PHOTOVOLTAIC POWERED PUMPING SYSTEMS: A CASE STUDY OF SOLLA TOGO

Access to improved potable water sources is recognized as one of the key factors in improving health and alleviating global poverty. In recently years, substantial investments have been made internationally in potable water infrastructure projects, allowing 2.3 billion people to gain access to potable water from 1990-2012. One such project was planned and installed in Solla, Togo, a rural village in the northern part of the country, from 2010-2012. Ethnographic studies revealed that, while the community has access to potable water, an estimated 45% of the village’s 1500 residents still rely on unprotected sources for drinking and cooking. Additionally, inequality in system use based on income level was revealed, with the higher income groups accessing the system more regularly than lower income groups. Cost, as well as the availability of cheaper sources, was identified as the main deterrent from using the new water distribution system. A new water-pricing scheme is investigated here with the intention of making the system accessible to a greater percentage of the population. Since 2012, a village-level water committee has been responsible for operations and maintenance (O&M), fulfilling the community management model that is recommended by many development theorists in order to create sustainable projects. The water committee received post-construction support, mostly in the form of technical support during system breakdowns, from the Togolese Ministry of Water and Sanitation (MWSVH). While this support has been valuable in maintaining a functional water supply system in Solla, the water committee still has managerial challenges, particularly with billing and fee collection. As a result, the water committee has only received 2% - 25% of the fees owed at each private connection and public tap stand, making their finances vulnerable when future repairs and capital replacements are necessary. A new management structure is proposed by the MWSVH that will pay utilities workers a wage and will hire an accountant in order to improve the local management and increase revenue. This proposal is analyzed under the new water pricing schemes that are presented. Initially, the rural water supply system was powered by a diesel-generator, but in 2013, a solar photo-voltaic power supply was installed. The new system proved a fiscal improvement for the village water committee, since it drastically reduced their annual O&M costs. However, the new system pumps a smaller volume of water on a daily basis and did not meet the community’s water needs during the dry season of 2014. A hydraulic network model was developed to investigate the system’s reliability under diesel-generator (DGPS) and solar photovoltaic (PVPS) power supplies. Additionally, a new system layout is proposed for the PVPS that allows pumping directly into the distribution line, circumventing the high head associated with pumping solely to the storage tank. It was determined that this new layout would allow for a greater volume of water to be provided to the demand points over the course of a day, meeting a greater fraction of the demand than with the current layout.

Leveraging Symbiotic Relationships for Emulation of Computer Networks

The lack of analytical models that can accurately describe large-scale networked systems makes empirical experimentation indispensable for understanding complex behaviors. Research on network testbeds for testing network protocols and distributed services, including physical, emulated, and federated testbeds, has made steady progress. Although the success of these testbeds is undeniable, they fail to provide: 1) scalability, for handling large-scale networks with hundreds or thousands of hosts and routers organized in different scenarios, 2) flexibility, for testing new protocols or applications in diverse settings, and 3) inter-operability, for combining simulated and real network entities in experiments. This dissertation tackles these issues in three different dimensions. First, we present SVEET, a system that enables inter-operability between real and simulated hosts. In order to increase the scalability of networks under study, SVEET enables time-dilated synchronization between real hosts and the discrete-event simulator. Realistic TCP congestion control algorithms are implemented in the simulator to allow seamless interactions between real and simulated hosts. SVEET is validated via extensive experiments and its capabilities are assessed through case studies involving real applications. Second, we present PrimoGENI, a system that allows a distributed discrete-event simulator, running in real-time, to interact with real network entities in a federated environment. PrimoGENI greatly enhances the flexibility of network experiments, through which a great variety of network conditions can be reproduced to examine what-if questions. Furthermore, PrimoGENI performs resource management functions, on behalf of the user, for instantiating network experiments on shared infrastructures. Finally, to further increase the scalability of network testbeds to handle large-scale high-capacity networks, we present a novel symbiotic simulation approach. We present SymbioSim, a testbed for large-scale network experimentation where a high-performance simulation system closely cooperates with an emulation system in a mutually beneficial way. On the one hand, the simulation system benefits from incorporating the traffic metadata from real applications in the emulation system to reproduce the realistic traffic conditions. On the other hand, the emulation system benefits from receiving the continuous updates from the simulation system to calibrate the traffic between real applications. Specific techniques that support the symbiotic approach include: 1) a model downscaling scheme that can significantly reduce the complexity of the large-scale simulation model, resulting in an efficient emulation system for modulating the high-capacity network traffic between real applications; 2) a queuing network model for the downscaled emulation system to accurately represent the network effects of the simulated traffic; and 3) techniques for reducing the synchronization overhead between the simulation and emulation systems.

Urban air mobility: modelling the future transport system

The advances in the aviation field, particularly the development of electric flying vehicles, as UAV and eVTOL, paved the way for setting Urban Air Mobility (UAM) services. UAM would provide services for passengers, goods and emergencies and could offer faster trips than ground ones. It is expected that early UAM operations will be performed at Very Low-Level airspace as 0-500 m Above Ground Level. The purpose of this research is to both explore the main features of UAM and test an aerial network model, which could be integrated in a multimodal transport system where ground and aerial mobility services are provided. Analyses on UAM transport system involved two sub-systems: the transport demand sub-system, i.e., the mobility requirements, and the transport supply sub-system, i.e., the service and facilities enabling mobility. At first, the UAM demand levels and features for an Airport Shuttle service have been explored through a suitable survey, by combining Revealed and Stated Preference methodologies, and by calibrating some discrete mode choice models. Then, the focus has been on the transport supply model for UAM services, by focusing on both the ground access points (vertiports) and the aerial network model. A suitable three-dimensional urban aerial network (3D-UAN) model that could support fast aerial connections between O/D pairs has been proposed. Some tests have been implemented to verify the feasibility of the proposed model. Some flying vehicles supporting an Airport Shuttle service have been simulated on the aerial network, which has been specified in terms of both topological features and link transport costs. The preliminary results have showed that the proposed 3D-UAN model could be suitable for supporting UAM services. As for transport engineering, the UAM system framework proposed in this thesis paves the way for further research on air-ground multimodality in urban areas.

Deep Learning applicato alla diagnostica dei trasformatori

Il trasformatore è uno degli elementi più importanti di una rete di trasmissione; essendo il tramite fra reti di alta e media tensione, il suo corretto funzionamento garantisce l’alimentazione di tutti i dispositivi e carichi connessi alla linea. Oltre a questo, il trasformatore è anche l’elemento più costoso di tutta la linea elettrica; la sua manutenzione è di vitale importanza per evitare costi elevati per la sostituzione e disagi lungo la linea. Qui entra in gioco il ruolo della diagnostica; attraverso misure periodiche e mirate sul trasformatore è possibile agire tempestivamente ed evitare tutti i fenomeni precedentemente elencati. Nell’elaborato si tratterà l’analisi del trasformatore elettrico trifase durante il suo funzionamento, evidenziando i sottocomponenti e le rispettive criticità; inoltre, verranno mostrate le varie tecniche di diagnostica del trasformatore, in modo tale da poter estrarre un indice legato allo stato di vita, ossia l’Health Index. Ad oggi esistono diverse tecniche di approccio al calcolo dell’Health Index, quella che viene presentata è una tecnica del tutto innovativa, ossia sviluppare una rete neurale artificiale (Artificial Neural Network, ANN) in grado di prevedere lo stato del trasformatore basandosi su misure effettuate sullo stesso. Dunque, verranno presentante le basi per lo sviluppo di una rete neurale, partendo dall’analisi e formattazione dei dati, fino alla fase di ottimizzazione delle prestazioni. Infine, si attraverseranno tutte le fasi intermedie di realizzazione del progetto da cui l’elaborato prende il titolo; osservando l’evoluzione di una rete neurale che si trasforma da un programma scritto in ambiente Python a una applicazione pronta all’uso per gli operatori durante le operazioni di diagnostica.

Machine Learning algorithm on seat comfort level measurement, mapping pressure values with body regions

The comfort level of the seat has a major effect on the usage of a vehicle; thus, car manufacturers have been working on elevating car seat comfort as much as possible. However, still, the testing and evaluation of comfort are done using exhaustive trial and error testing and evaluation of data. In this thesis, we resort to machine learning and Artificial Neural Networks (ANN) to develop a fully automated approach. Even though this approach has its advantages in minimizing time and using a large set of data, it takes away the degree of freedom of the engineer on making decisions. The focus of this study is on filling the gap in a two-step comfort level evaluation which used pressure mapping with body regions to evaluate the average pressure supported by specific body parts and the Self-Assessment Exam (SAE) questions on evaluation of the person’s interest. This study has created a machine learning algorithm that works on giving a degree of freedom to the engineer in making a decision when mapping pressure values with body regions using ANN. The mapping is done with 92% accuracy and with the help of a Graphical User Interface (GUI) that facilitates the process during the testing time of comfort level evaluation of the car seat, which decreases the duration of the test analysis from days to hours.

Application of active and transfer learning to wound image segmentation

Wound management is a fundamental task in standard clinical practice. Automated solutions already exist for humans, but there is a lack of applications on wound management for pets. The importance of a precise and efficient wound assessment is helpful to improve diagnosis and to increase the effectiveness of treatment plans for the chronic wounds. The goal of the research was to propose an automated pipeline capable of segmenting natural light-reflected wound images of animals. Two datasets composed by light-reflected images were used in this work: Deepskin dataset, 1564 human wound images obtained during routine dermatological exams, with 145 manual annotated images; Petwound dataset, a set of 290 wound photos of dogs and cats with 0 annotated images. Two implementations of U-Net Convolutioal Neural Network model were proposed for the automated segmentation. Active Semi-Supervised Learning techniques were applied for human-wound images to perform segmentation from 10% of annotated images. Then the same models were trained, via Transfer Learning, adopting an Active Semi- upervised Learning to unlabelled animal-wound images. The combination of the two training strategies proved their effectiveness in generating large amounts of annotated samples (94% of Deepskin, 80% of PetWound) with the minimal human intervention. The correctness of automated segmentation were evaluated by clinical experts at each round of training thus we can assert that the results obtained in this thesis stands as a reliable solution to perform a correct wound image segmentation. The use of Transfer Learning and Active Semi-Supervied Learning allows to minimize labelling effort from clinicians, even requiring no starting manual annotation at all. Moreover the performances of the model with limited number of parameters suggest the implementation of smartphone-based application to this topic, helping the future standardization of light-reflected images as acknowledge medical images.

Artificial evolution on network structures : how time and space influence dynamics

Abstract The main objective of this work is to show how the choice of the temporal dimension and of the spatial structure of the population influences an artificial evolutionary process. In the field of Artificial Evolution we can observe a common trend in synchronously evolv¬ing panmictic populations, i.e., populations in which any individual can be recombined with any other individual. Already in the '90s, the works of Spiessens and Manderick, Sarma and De Jong, and Gorges-Schleuter have pointed out that, if a population is struc¬tured according to a mono- or bi-dimensional regular lattice, the evolutionary process shows a different dynamic with respect to the panmictic case. In particular, Sarma and De Jong have studied the selection pressure (i.e., the diffusion of a best individual when the only selection operator is active) induced by a regular bi-dimensional structure of the population, proposing a logistic modeling of the selection pressure curves. This model supposes that the diffusion of a best individual in a population follows an exponential law. We show that such a model is inadequate to describe the process, since the growth speed must be quadratic or sub-quadratic in the case of a bi-dimensional regular lattice. New linear and sub-quadratic models are proposed for modeling the selection pressure curves in, respectively, mono- and bi-dimensional regu¬lar structures. These models are extended to describe the process when asynchronous evolutions are employed. Different dynamics of the populations imply different search strategies of the resulting algorithm, when the evolutionary process is used to solve optimisation problems. A benchmark of both discrete and continuous test problems is used to study the search characteristics of the different topologies and updates of the populations. In the last decade, the pioneering studies of Watts and Strogatz have shown that most real networks, both in the biological and sociological worlds as well as in man-made structures, have mathematical properties that set them apart from regular and random structures. In particular, they introduced the concepts of small-world graphs, and they showed that this new family of structures has interesting computing capabilities. Populations structured according to these new topologies are proposed, and their evolutionary dynamics are studied and modeled. We also propose asynchronous evolutions for these structures, and the resulting evolutionary behaviors are investigated. Many man-made networks have grown, and are still growing incrementally, and explanations have been proposed for their actual shape, such as Albert and Barabasi's preferential attachment growth rule. However, many actual networks seem to have undergone some kind of Darwinian variation and selection. Thus, how these networks might have come to be selected is an interesting yet unanswered question. In the last part of this work, we show how a simple evolutionary algorithm can enable the emrgence o these kinds of structures for two prototypical problems of the automata networks world, the majority classification and the synchronisation problems. Synopsis L'objectif principal de ce travail est de montrer l'influence du choix de la dimension temporelle et de la structure spatiale d'une population sur un processus évolutionnaire artificiel. Dans le domaine de l'Evolution Artificielle on peut observer une tendence à évoluer d'une façon synchrone des populations panmictiques, où chaque individu peut être récombiné avec tout autre individu dans la population. Déjà dans les année '90, Spiessens et Manderick, Sarma et De Jong, et Gorges-Schleuter ont observé que, si une population possède une structure régulière mono- ou bi-dimensionnelle, le processus évolutionnaire montre une dynamique différente de celle d'une population panmictique. En particulier, Sarma et De Jong ont étudié la pression de sélection (c-à-d la diffusion d'un individu optimal quand seul l'opérateur de sélection est actif) induite par une structure régulière bi-dimensionnelle de la population, proposant une modélisation logistique des courbes de pression de sélection. Ce modèle suppose que la diffusion d'un individu optimal suit une loi exponentielle. On montre que ce modèle est inadéquat pour décrire ce phénomène, étant donné que la vitesse de croissance doit obéir à une loi quadratique ou sous-quadratique dans le cas d'une structure régulière bi-dimensionnelle. De nouveaux modèles linéaires et sous-quadratique sont proposés pour des structures mono- et bi-dimensionnelles. Ces modèles sont étendus pour décrire des processus évolutionnaires asynchrones. Différentes dynamiques de la population impliquent strategies différentes de recherche de l'algorithme résultant lorsque le processus évolutionnaire est utilisé pour résoudre des problèmes d'optimisation. Un ensemble de problèmes discrets et continus est utilisé pour étudier les charactéristiques de recherche des différentes topologies et mises à jour des populations. Ces dernières années, les études de Watts et Strogatz ont montré que beaucoup de réseaux, aussi bien dans les mondes biologiques et sociologiques que dans les structures produites par l'homme, ont des propriétés mathématiques qui les séparent à la fois des structures régulières et des structures aléatoires. En particulier, ils ont introduit la notion de graphe sm,all-world et ont montré que cette nouvelle famille de structures possède des intéressantes propriétés dynamiques. Des populations ayant ces nouvelles topologies sont proposés, et leurs dynamiques évolutionnaires sont étudiées et modélisées. Pour des populations ayant ces structures, des méthodes d'évolution asynchrone sont proposées, et la dynamique résultante est étudiée. Beaucoup de réseaux produits par l'homme se sont formés d'une façon incrémentale, et des explications pour leur forme actuelle ont été proposées, comme le preferential attachment de Albert et Barabàsi. Toutefois, beaucoup de réseaux existants doivent être le produit d'un processus de variation et sélection darwiniennes. Ainsi, la façon dont ces structures ont pu être sélectionnées est une question intéressante restée sans réponse. Dans la dernière partie de ce travail, on montre comment un simple processus évolutif artificiel permet à ce type de topologies d'émerger dans le cas de deux problèmes prototypiques des réseaux d'automates, les tâches de densité et de synchronisation.

The use of artificial neural networks in analysing the nutritional ecology of Chrysomya megacephala (F.) (Diptera: Calliphoridae), compared with a statistical model

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)