A nested Petri net framework for modeling and analyzing multi-agent systems

In the past two decades, multi-agent systems (MAS) have emerged as a new paradigm for conceptualizing large and complex distributed software systems. A multi-agent system view provides a natural abstraction for both the structure and the behavior of modern-day software systems. Although there were many conceptual frameworks for using multi-agent systems, there was no well established and widely accepted method for modeling multi-agent systems. This dissertation research addressed the representation and analysis of multi-agent systems based on model-oriented formal methods. The objective was to provide a systematic approach for studying MAS at an early stage of system development to ensure the quality of design. ^ Given that there was no well-defined formal model directly supporting agent-oriented modeling, this study was centered on three main topics: (1) adapting a well-known formal model, predicate transition nets (PrT nets), to support MAS modeling; (2) formulating a modeling methodology to ease the construction of formal MAS models; and (3) developing a technique to support machine analysis of formal MAS models using model checking technology. PrT nets were extended to include the notions of dynamic structure, agent communication and coordination to support agent-oriented modeling. An aspect-oriented technique was developed to address the modularity of agent models and compositionality of incremental analysis. A set of translation rules were defined to systematically translate formal MAS models to concrete models that can be verified through the model checker SPIN (Simple Promela Interpreter). ^ This dissertation presents the framework developed for modeling and analyzing MAS, including a well-defined process model based on nested PrT nets, and a comprehensive methodology to guide the construction and analysis of formal MAS models.^

Evaluation of wind-induced internal pressure in low-rise buildings: A multi scale experimental and numerical approach

Hurricane is one of the most destructive and costly natural hazard to the built environment and its impact on low-rise buildings, particularity, is beyond acceptable. The major objective of this research was to perform a parametric evaluation of internal pressure (IP) for wind-resistant design of low-rise buildings and wind-driven natural ventilation applications. For this purpose, a multi-scale experimental, i.e. full-scale at Wall of Wind (WoW) and small-scale at Boundary Layer Wind Tunnel (BLWT), and a Computational Fluid Dynamics (CFD) approach was adopted. This provided new capability to assess wind pressures realistically on internal volumes ranging from small spaces formed between roof tiles and its deck to attic to room partitions. Effects of sudden breaching, existing dominant openings on building envelopes as well as compartmentalization of building interior on the IP were systematically investigated. Results of this research indicated: (i) for sudden breaching of dominant openings, the transient overshooting response was lower than the subsequent steady state peak IP and internal volume correction for low-wind-speed testing facilities was necessary. For example a building without volume correction experienced a response four times faster and exhibited 30–40% lower mean and peak IP; (ii) for existing openings, vent openings uniformly distributed along the roof alleviated, whereas one sided openings aggravated the IP; (iii) larger dominant openings exhibited a higher IP on the building envelope, and an off-center opening on the wall exhibited (30–40%) higher IP than center located openings; (iv) compartmentalization amplified the intensity of IP and; (v) significant underneath pressure was measured for field tiles, warranting its consideration during net pressure evaluations. The study aimed at wind driven natural ventilation indicated: (i) the IP due to cross ventilation was 1.5 to 2.5 times higher for Ainlet/Aoutlet>1 compared to cases where Ainlet/Aoutlet<1, this in effect reduced the mixing of air inside the building and hence the ventilation effectiveness; (ii) the presence of multi-room partitioning increased the pressure differential and consequently the air exchange rate. Overall good agreement was found between the observed large-scale, small-scale and CFD based IP responses. Comparisons with ASCE 7-10 consistently demonstrated that the code underestimated peak positive and suction IP.

Multi-Tissue Stable Isotope Analysis and Acoustic Telemetry Reveal Seasonal Variability in the Trophic Interactions of Juvenile Bull Sharks in a Coastal Estuary

Understanding how natural and anthropogenic drivers affect extant food webs is critical to predicting the impacts of climate change and habitat alterations on ecosystem dynamics. In the Florida Everglades, seasonal reductions in freshwater flow and precipitation lead to annual migrations of aquatic taxa from marsh habitats to deep-water refugia in estuaries. The timing and intensity of freshwater reductions, however, will be modified by ongoing ecosystem restoration and predicted climate change. Understanding the importance of seasonally pulsed resources to predators is critical to predicting the impacts of management and climate change on their populations. As with many large predators, however, it is difficult to determine to what extent predators like bull sharks (Carcharhinus leucas) in the coastal Everglades make use of prey pulses currently. We used passive acoustic telemetry to determine whether shark movements responded to the pulse of marsh prey. To investigate the possibility that sharks fed on marsh prey, we modelled the predicted dynamics of stable isotope values in bull shark blood and plasma under different assumptions of temporal variability in shark diets and physiological dynamics of tissue turnover and isotopic discrimination. Bull sharks increased their use of upstream channels during the late dry season, and although our previous work shows long-term specialization in the diets of sharks, stable isotope values suggested that some individuals adjusted their diets to take advantage of prey entering the system from the marsh, and as such this may be an important resource for the nursery. Restoration efforts are predicted to increase hydroperiods and marsh water levels, likely shifting the timing, duration and intensity of prey pulses, which could have negative consequences for the bull shark population and/or induce shifts in behaviour. Understanding the factors influencing the propensity to specialize or adopt more flexible trophic interactions will be an important step in fully understanding the ecological role of predators and how ecological roles may vary with environmental and anthropogenic changes.

Multi-year weather data in the remote semi-arid to arid mountain region of the Saharan flank of the Atlas Mountains

In 2001, a weather and climate monitoring network was established along the temperature and aridity gradient between the sub-humid Moroccan High Atlas Mountains and the former end lake of the Middle Drâa in a pre-Saharan environment. The highest Automated Weather Stations (AWS) was installed just below the M'Goun summit at 3850 m, the lowest station Lac Iriki was at 450 m. This network of 13 AWS stations was funded and maintained by the German IMPETUS (BMBF Grant 01LW06001A, North Rhine-Westphalia Grant 313-21200200) project and since 2011 five stations were further maintained by the GERMAN DFG Fennec project (FI 786/3-1), this way some stations of the AWS network provided data for almost 12 years from 2001-2012. Standard meteorological variables such as temperature, humidity, and wind were measured at an altitude of 2 m above ground. Other meteorological variables comprise precipitation, station pressure, solar irradiance, soil temperature at different depths and for high mountain station snow water equivalent. The stations produced data summaries for 5-minute-precipitation-data, 10- or 15-minute-data and a daily summary of all other variables. This network is a unique resource of multi-year weather data in the remote semi-arid to arid mountain region of the Saharan flank of the Atlas Mountains. The network is described in Schulz et al. (2010) and its further continuation until 2012 is briefly discussed in Redl et al. (2015, doi:10.1175/MWR-D-15-0223.1) and Redl et al. (2016, doi:10.1002/2015JD024443).

AdaptMCloud: uma estratégia para adaptação dinâmica de aplicações Multi-Cloud

Multi-Cloud Applications are composed of services offered by multiple cloud platforms where the user/developer has full knowledge of the use of such platforms. The use of multiple cloud platforms avoids the following problems: (i) vendor lock-in, which is dependency on the application of a certain cloud platform, which is prejudicial in the case of degradation or failure of platform services, or even price increasing on service usage; (ii) degradation or failure of the application due to fluctuations in quality of service (QoS) provided by some cloud platform, or even due to a failure of any service. In multi-cloud scenario is possible to change a service in failure or with QoS problems for an equivalent of another cloud platform. So that an application can adopt the perspective multi-cloud is necessary to create mechanisms that are able to select which cloud services/platforms should be used in accordance with the requirements determined by the programmer/user. In this context, the major challenges in terms of development of such applications include questions such as: (i) the choice of which underlying services and cloud computing platforms should be used based on the defined user requirements in terms of functionality and quality (ii) the need to continually monitor the dynamic information (such as response time, availability, price, availability), related to cloud services, in addition to the wide variety of services, and (iii) the need to adapt the application if QoS violations affect user defined requirements. This PhD thesis proposes an approach for dynamic adaptation of multi-cloud applications to be applied when a service is unavailable or when the requirements set by the user/developer point out that other available multi-cloud configuration meets more efficiently. Thus, this work proposes a strategy composed of two phases. The first phase consists of the application modeling, exploring the similarities representation capacity and variability proposals in the context of the paradigm of Software Product Lines (SPL). In this phase it is used an extended feature model to specify the cloud service configuration to be used by the application (similarities) and the different possible providers for each service (variability). Furthermore, the non-functional requirements associated with cloud services are specified by properties in this model by describing dynamic information about these services. The second phase consists of an autonomic process based on MAPE-K control loop, which is responsible for selecting, optimally, a multicloud configuration that meets the established requirements, and perform the adaptation. The adaptation strategy proposed is independent of the used programming technique for performing the adaptation. In this work we implement the adaptation strategy using various programming techniques such as aspect-oriented programming, context-oriented programming and components and services oriented programming. Based on the proposed steps, we tried to assess the following: (i) the process of modeling and the specification of non-functional requirements can ensure effective monitoring of user satisfaction; (ii) if the optimal selection process presents significant gains compared to sequential approach; and (iii) which techniques have the best trade-off when compared efforts to development/modularity and performance.

An application of evolutionary algorithms for WAG optimisation in the Norne Field

Water-alternating-gas (WAG) is an enhanced oil recovery method combining the improved macroscopic sweep of water flooding with the improved microscopic displacement of gas injection. The optimal design of the WAG parameters is usually based on numerical reservoir simulation via trial and error, limited by the reservoir engineer’s availability. Employing optimisation techniques can guide the simulation runs and reduce the number of function evaluations. In this study, robust evolutionary algorithms are utilized to optimise hydrocarbon WAG performance in the E-segment of the Norne field. The first objective function is selected to be the net present value (NPV) and two global semi-random search strategies, a genetic algorithm (GA) and particle swarm optimisation (PSO) are tested on different case studies with different numbers of controlling variables which are sampled from the set of water and gas injection rates, bottom-hole pressures of the oil production wells, cycle ratio, cycle time, the composition of the injected hydrocarbon gas (miscible/immiscible WAG) and the total WAG period. In progressive experiments, the number of decision-making variables is increased, increasing the problem complexity while potentially improving the efficacy of the WAG process. The second objective function is selected to be the incremental recovery factor (IRF) within a fixed total WAG simulation time and it is optimised using the same optimisation algorithms. The results from the two optimisation techniques are analyzed and their performance, convergence speed and the quality of the optimal solutions found by the algorithms in multiple trials are compared for each experiment. The distinctions between the optimal WAG parameters resulting from NPV and oil recovery optimisation are also examined. This is the first known work optimising over this complete set of WAG variables. The first use of PSO to optimise a WAG project at the field scale is also illustrated. Compared to the reference cases, the best overall values of the objective functions found by GA and PSO were 13.8% and 14.2% higher, respectively, if NPV is optimised over all the above variables, and 14.2% and 16.2% higher, respectively, if IRF is optimised.

Algorithme de branch-and-price-and-cut pour le problème de conception de réseaux avec coûts fixes, capacités et un seul produit

De nombreux problèmes liés aux domaines du transport, des télécommunications et de la logistique peuvent être modélisés comme des problèmes de conception de réseaux. Le problème classique consiste à transporter un flot (données, personnes, produits, etc.) sur un réseau sous un certain nombre de contraintes dans le but de satisfaire la demande, tout en minimisant les coûts. Dans ce mémoire, on se propose d'étudier le problème de conception de réseaux avec coûts fixes, capacités et un seul produit, qu'on transforme en un problème équivalent à plusieurs produits de façon à améliorer la valeur de la borne inférieure provenant de la relaxation continue du modèle. La méthode que nous présentons pour la résolution de ce problème est une méthode exacte de branch-and-price-and-cut avec une condition d'arrêt, dans laquelle nous exploitons à la fois la méthode de génération de colonnes, la méthode de génération de coupes et l'algorithme de branch-and-bound. Ces méthodes figurent parmi les techniques les plus utilisées en programmation linéaire en nombres entiers. Nous testons notre méthode sur deux groupes d'instances de tailles différentes (gran-des et très grandes), et nous la comparons avec les résultats donnés par CPLEX, un des meilleurs logiciels permettant de résoudre des problèmes d'optimisation mathématique, ainsi qu’avec une méthode de branch-and-cut. Il s'est avéré que notre méthode est prometteuse et peut donner de bons résultats, en particulier pour les instances de très grandes tailles.

Person Re-Identification using Deep Convnets with Multi-task Learning

Person re-identification involves recognizing a person across non-overlapping camera views, with different pose, illumination, and camera characteristics. We propose to tackle this problem by training a deep convolutional network to represent a person’s appearance as a low-dimensional feature vector that is invariant to common appearance variations encountered in the re-identification problem. Specifically, a Siamese-network architecture is used to train a feature extraction network using pairs of similar and dissimilar images. We show that use of a novel multi-task learning objective is crucial for regularizing the network parameters in order to prevent over-fitting due to the small size the training dataset. We complement the verification task, which is at the heart of re-identification, by training the network to jointly perform verification, identification, and to recognise attributes related to the clothing and pose of the person in each image. Additionally, we show that our proposed approach performs well even in the challenging cross-dataset scenario, which may better reflect real-world expected performance. 

Automatically Evolving Rotation-invariant Texture Image Descriptors by Genetic Programming

In computer vision, training a model that performs classification effectively is highly dependent on the extracted features, and the number of training instances. Conventionally, feature detection and extraction are performed by a domain-expert who, in many cases, is expensive to employ and hard to find. Therefore, image descriptors have emerged to automate these tasks. However, designing an image descriptor still requires domain-expert intervention. Moreover, the majority of machine learning algorithms require a large number of training examples to perform well. However, labelled data is not always available or easy to acquire, and dealing with a large dataset can dramatically slow down the training process. In this paper, we propose a novel Genetic Programming based method that automatically synthesises a descriptor using only two training instances per class. The proposed method combines arithmetic operators to evolve a model that takes an image and generates a feature vector. The performance of the proposed method is assessed using six datasets for texture classification with different degrees of rotation, and is compared with seven domain-expert designed descriptors. The results show that the proposed method is robust to rotation, and has significantly outperformed, or achieved a comparable performance to, the baseline methods.

Climbing Up the Leaderboard: An Empirical Study of Applying Gamification Techniques to a Computer Programming Class

Conventional taught learning practices often experience difficulties in keeping students motivated and engaged. Video games, however, are very successful at sustaining high levels of motivation and engagement through a set of tasks for hours without apparent loss of focus. In addition, gamers solve complex problems within a gaming environment without feeling fatigue or frustration, as they would typically do with a comparable learning task. Based on this notion, the academic community is keen on exploring methods that can deliver deep learner engagement and has shown increased interest in adopting gamification – the integration of gaming elements, mechanics, and frameworks into non-game situations and scenarios – as a means to increase student engagement and improve information retention. Its effectiveness when applied to education has been debatable though, as attempts have generally been restricted to one-dimensional approaches such as transposing a trivial reward system onto existing teaching materials and/or assessments. Nevertheless, a gamified, multi-dimensional, problem-based learning approach can yield improved results even when applied to a very complex and traditionally dry task like the teaching of computer programming, as shown in this paper. The presented quasi-experimental study used a combination of instructor feedback, real time sequence of scored quizzes, and live coding to deliver a fully interactive learning experience. More specifically, the “Kahoot!” Classroom Response System (CRS), the classroom version of the TV game show “Who Wants To Be A Millionaire?”, and Codecademy’s interactive platform formed the basis for a learning model which was applied to an entry-level Python programming course. Students were thus allowed to experience multiple interlocking methods similar to those commonly found in a top quality game experience. To assess gamification’s impact on learning, empirical data from the gamified group were compared to those from a control group who was taught through a traditional learning approach, similar to the one which had been used during previous cohorts. Despite this being a relatively small-scale study, the results and findings for a number of key metrics, including attendance, downloading of course material, and final grades, were encouraging and proved that the gamified approach was motivating and enriching for both students and instructors.

Building Better Nurse Scheduling Algorithms

The aim of this research is twofold: Firstly, to model and solve a complex nurse scheduling problem with an integer programming formulation and evolutionary algorithms. Secondly, to detail a novel statistical method of comparing and hence building better scheduling algorithms by identifying successful algorithm modifications. The comparison method captures the results of algorithms in a single figure that can then be compared using traditional statistical techniques. Thus, the proposed method of comparing algorithms is an objective procedure designed to assist in the process of improving an algorithm. This is achieved even when some results are non-numeric or missing due to infeasibility. The final algorithm outperforms all previous evolutionary algorithms, which relied on human expertise for modification.

GPRM: a high performance programming framework for manycore processors

Processors with large numbers of cores are becoming commonplace. In order to utilise the available resources in such systems, the programming paradigm has to move towards increased parallelism. However, increased parallelism does not necessarily lead to better performance. Parallel programming models have to provide not only flexible ways of defining parallel tasks, but also efficient methods to manage the created tasks. Moreover, in a general-purpose system, applications residing in the system compete for the shared resources. Thread and task scheduling in such a multiprogrammed multithreaded environment is a significant challenge. In this thesis, we introduce a new task-based parallel reduction model, called the Glasgow Parallel Reduction Machine (GPRM). Our main objective is to provide high performance while maintaining ease of programming. GPRM supports native parallelism; it provides a modular way of expressing parallel tasks and the communication patterns between them. Compiling a GPRM program results in an Intermediate Representation (IR) containing useful information about tasks, their dependencies, as well as the initial mapping information. This compile-time information helps reduce the overhead of runtime task scheduling and is key to high performance. Generally speaking, the granularity and the number of tasks are major factors in achieving high performance. These factors are even more important in the case of GPRM, as it is highly dependent on tasks, rather than threads. We use three basic benchmarks to provide a detailed comparison of GPRM with Intel OpenMP, Cilk Plus, and Threading Building Blocks (TBB) on the Intel Xeon Phi, and with GNU OpenMP on the Tilera TILEPro64. GPRM shows superior performance in almost all cases, only by controlling the number of tasks. GPRM also provides a low-overhead mechanism, called “Global Sharing”, which improves performance in multiprogramming situations. We use OpenMP, as the most popular model for shared-memory parallel programming as the main GPRM competitor for solving three well-known problems on both platforms: LU factorisation of Sparse Matrices, Image Convolution, and Linked List Processing. We focus on proposing solutions that best fit into the GPRM’s model of execution. GPRM outperforms OpenMP in all cases on the TILEPro64. On the Xeon Phi, our solution for the LU Factorisation results in notable performance improvement for sparse matrices with large numbers of small blocks. We investigate the overhead of GPRM’s task creation and distribution for very short computations using the Image Convolution benchmark. We show that this overhead can be mitigated by combining smaller tasks into larger ones. As a result, GPRM can outperform OpenMP for convolving large 2D matrices on the Xeon Phi. Finally, we demonstrate that our parallel worksharing construct provides an efficient solution for Linked List processing and performs better than OpenMP implementations on the Xeon Phi. The results are very promising, as they verify that our parallel programming framework for manycore processors is flexible and scalable, and can provide high performance without sacrificing productivity.

Analysis of the moisture effect on the mechanical behaviour of an adhesively bonded joint under proportional multi-axial loads

The objective of the study is to identify the 3D behaviour of an adhesive in an assembly, and to take into account the effect of ageing in a marine environment. To that end, three different tests were employed. Gravimetric analyses were used to determine the water diffusion kinetics in the adhesive. Bulk tensile tests were performed to highlight the effects of humid ageing on the adhesive behaviour. Modified Arcan tests were performed for several ageing times to obtain the experimental database which was necessary to identify constitutive models. A Mahnken-Schlimmer type model was determined for the unaged state according to a procedure developed in a previous study. This identification used inverse techniques. It was based on the unaged modified Arcan results and on a coupling between an optimisation routine and finite-element analysis. Then, a global inverse identification procedure was developed. Its aim was to relate the unaged parameters to the moisture concentration and overcome the difficulties usually associated with ageing of bonded assemblies in a humid environment: a non-uniformity of the stress state and a gradient of mechanical properties in the adhesive. This procedure was similar to the one used in the first part but needed modified Arcan results for several ageing times. It also required an initial assumption for the evolution of the Mahnken-Schlimmer parameters with the moisture concentration.

Building Better Nurse Scheduling Algorithms

The aim of this research is twofold: Firstly, to model and solve a complex nurse scheduling problem with an integer programming formulation and evolutionary algorithms. Secondly, to detail a novel statistical method of comparing and hence building better scheduling algorithms by identifying successful algorithm modifications. The comparison method captures the results of algorithms in a single figure that can then be compared using traditional statistical techniques. Thus, the proposed method of comparing algorithms is an objective procedure designed to assist in the process of improving an algorithm. This is achieved even when some results are non-numeric or missing due to infeasibility. The final algorithm outperforms all previous evolutionary algorithms, which relied on human expertise for modification.

Building Better Nurse Scheduling Algorithms

The aim of this research is twofold: Firstly, to model and solve a complex nurse scheduling problem with an integer programming formulation and evolutionary algorithms. Secondly, to detail a novel statistical method of comparing and hence building better scheduling algorithms by identifying successful algorithm modifications. The comparison method captures the results of algorithms in a single figure that can then be compared using traditional statistical techniques. Thus, the proposed method of comparing algorithms is an objective procedure designed to assist in the process of improving an algorithm. This is achieved even when some results are non-numeric or missing due to infeasibility. The final algorithm outperforms all previous evolutionary algorithms, which relied on human expertise for modification.