Submitted by Alison Vanceto (alison-vanceto@hotmail.com) on 2017-05-09T12:58:24Z No. of bitstreams: 1 DissBRS.pdf: 3255630 bytes, checksum: e4d94f1ee923813cfd91e76e992dbfa8 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-06-01T18:49:57Z (GMT) No. of bitstreams: 1 DissBRS.pdf: 3255630 bytes, checksum: e4d94f1ee923813cfd91e76e992dbfa8 (MD5) / Approved for entry into archive by Ronildo Prado (ronisp@ufscar.br) on 2017-06-01T18:50:12Z (GMT) No. of bitstreams: 1 DissBRS.pdf: 3255630 bytes, checksum: e4d94f1ee923813cfd91e76e992dbfa8 (MD5) / Made available in DSpace on 2017-06-01T19:49:40Z (GMT). No. of bitstreams: 1 DissBRS.pdf: 3255630 bytes, checksum: e4d94f1ee923813cfd91e76e992dbfa8 (MD5) Previous issue date: 2016-06-30 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / Context: Traditional testing approaches, considering the context of Adaptive Systems (ASs), have been evaluated incipiently and superficially with respect to their effectiveness in identifying faults. It is very clear the challenge of guaranteeing the correctness of ASs, due to the adaptive properties of these systems. Thus, the following question raises: how to characterise adequately and broadly the difficulties for testing ASs? There is not an approach that is driven by challenges found in the literature. Goals: In this work, we analysed and characterised the challenges faced by the testing activity when applied to ASSs. By characterising the challenges, we investigated testing strategies based on them. Methodology: In order to achieve the goals, we performed the following steps: (i) characterising the state-of-art of testing of ASs, using a Systematic Literature Review (SLR); (ii) investigating the challenges in source codes of real ASs that are available online in source code repositories; (iii) defining the testing strategy named T, based on the characterised challenges; (iv) defining the testing strategy T* based on the challenges and comprised by three testing approaches; (v) evaluating the effectiveness of the strategies T and T*; and (vi) investigating the challenges along the application of T and T*. Results: the main results were: (i) a characterisation of fine-grained and coarse-grained challenges for testing ASs; and (ii) two testing strategies based on such challenges. We evaluated the strategies by running an exploratory study which encompassed one AS and both testing strategies. Conclusion: The characterisation of challenges supported the definition and execution of the testing strategies. These strategies were able to reduce the impact of the challenges and to identify faults in the AS. / Contexto: Abordagens de teste tradicionais sao ainda incipientes e superficialmente avaliadas quanto a sua efetividade em revelar defeitos em Sistemas Adaptativos (SAs). E nítido o desafio de garantir a corretude de SAs levando-se em consideração as características de adaptacao desses sistemas, culminando na seguinte pergunta: como caracterizar adequadamente, de forma abrangente, as dificuldades para se testar SAs? Na literatura nao foi encontrada qualquer estrategia de teste que seja guiada por desafios de teste caracterizados. Objetivos: Neste trabalho buscou-se compreender e caracterizar os desafios impostos a atividade de teste de SAs. Uma vez caracterizados, os desafios, propos-se investigar estrategias de teste baseadas nestes desafios. Metodologia: Para atingir o objetivo seguiram-se as etapas: (i) caracterizacao do estado da arte por meio de uma Revisao Sistematica (RS); (ii) investigacao da existencia de desafios de teste em SAs presentes em repositórios de código-fonte; (iii) definho de uma estrategia de teste, denominada T, com base nos desafios caracterizados da RS; (iv) definicao de uma estrategia de teste combinada, denominada T*, composta por três abordagens de teste; (v) avaliaçao da efetividade das estrategias T e T*; e, por fim, (vi) investigacao da presença dos desafios nos SAs analisados, por meio das estrategias T e T*. Resultados: Os principais resultados foram: (i) uma caraterizacao de desafios para o teste de SAs; e (ii) e uma definicao de estrategias de teste baseadas nos desafios caracterizados. A avaliacao foi realizada com a aplicacao de um estudo exploratório, utilizando um SA, de uma estrategia de teste somente baseada em desafios (T) e de uma estrategia de teste baseada em desafios combinada com outras abordagens (T*), a fim de comparar os resultados de ambas as estrategias. Conclusão: A caracterizacao dos desafios auxiliou na definicao e execucao de estrategias de teste, as quais mitigaram os desafios presentes nos SAs e puderam identificar defeitos.

Submitted by Cláudia Bueno (claudiamoura18@gmail.com) on 2015-10-16T21:23:43Z No. of bitstreams: 2 Tese - Sergio Almeida Loiola - 2014.pdf: 8526084 bytes, checksum: 2948c2f6e11578a77db5880e24463e41 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2015-10-19T13:30:31Z (GMT) No. of bitstreams: 2 Tese - Sergio Almeida Loiola - 2014.pdf: 8526084 bytes, checksum: 2948c2f6e11578a77db5880e24463e41 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2015-10-19T13:30:31Z (GMT). No. of bitstreams: 2 Tese - Sergio Almeida Loiola - 2014.pdf: 8526084 bytes, checksum: 2948c2f6e11578a77db5880e24463e41 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2014-04-25 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This thesis proposes hypothesis about the evolution of complex adaptive systems in the context of the interaction between society and the paleoclimate variability, based on comparative studies of bibliographic sources available in the literature: articles, research reports and data. The research was oriented in one cognitive and evolutionary perspective, subsidized by the complex geographical approach, structured by the author. Founded on archaeological evidence, the thesis argues that complex adaptive systems developed after 40 thousand years ago are between the main factors which cooperated in the social and cultural complexity, fixing on the territory, organizing large groups, the continuous pursuit of innovation and acceleration of cultural evolution and cognitive, established before the Holocene, as well as diversification of similar adaptive strategies in different parts of the world during this period; case of systematic collect, cultivation, domestication and agriculture. Between complex adaptive systems developed before the Holocene is cited: the science, the belief system, art, spoken language, symbolic communication, and the socio-political organization to manage large groups. Transferred by cultural exchanges, complex adaptive systems migrated with people, giving great adaptive differential to their carriers to favorable conditions in interglacial Holocene: an increase of temperature, humidity, rainfall, CO² in the atmosphere from 180 ppm to 280 ppm, and expansion the diversity and productivity of the species. Under non- linear processes, the successive period of intensification of innovation were associated with severe glacial and interglacial transitions of the 26 Dansgaard-Oeschger and Heinrich events (D-O and H), during the Würm Glacial, especially from 40,000 years ago, after “great leap” of modern humans. Based on the successes and failures of societies, imprevisibility and frequency of abrupt paleoclimatic changes, we present a proposal for Adaptive Sustainability Panel, focusing on three scenarios: global warming (Plan A), global cooling (Plan B) and climatic instability (Plan C), in order to increase the potential for proactive resilience in societies. / Esta tese propõe hipótese acerca da evolução de sistemas complexos adaptativos no contexto da interação entre a sociedade e a variabilidade paleoclimática, pautada em estudos comparativos de fontes bibliográficas disponíveis na literatura: artigos, relatórios de pesquisa e dados. A pesquisa foi orientada numa perspectiva cognitiva e evolutiva, subsidiada pela abordagem geográfica complexa, estruturada pelo autor. A partir de evidências arqueológicas, a tese argumenta que sistemas complexos adaptativos desenvolvidos após 40 mil anos atrás estão entre os principais fatores que cooperaram para a complefixificação sociocultural, fixação no território, organização de grandes grupamentos, a continua busca de inovação e a aceleração da evolução cultural e cognitiva, estabelecidas antes do Holoceno, bem como a diversificação de estratégias adaptativas similares, em diferentes partes do globo nesse período; caso da coleta sistemática, do cultivo, da domesticação e da agricultura. Entre os sistemas complexos adaptativos desenvolvidos antes do Holoceno estiveram à ciência, o sistema de crenças, a arte, a linguagem falada, a comunicação simbólica, e a organização sociopolítica para gerir grandes grupos. Transferidos por trocas culturais, os sistemas complexos adaptativos migraram junto com os povos, conferindo grande diferencial adaptativo aos seus portadores às condições favoráveis do interglacial Holoceno: incremento da temperatura, da umidade, das chuvas, do CO² na atmosfera de 180 ppm para 280 ppm, e a ampliação da diversidade e a produtividade das espécies. Sob processos não lineares, os sucessivos período de intensificação da inovação estiveram correlacionados com severas transições glaciais e interglaciais dos 26 eventos Dansgaard-Oeschger e Heinrich (D-O e H), durante o Glacial Würm, especialmente a partir de 40 mil anos atrás, após o grande salto dos humanos modernos. Com base nos sucessos e fracassos das sociedades, a imponderabilidade e a frequência das alterações paleoclimáticas abruptas, apresentamos uma proposta para um Painel da Sustentabilidade Adaptativa, focando três cenários: de aquecimento global (Plano A), resfriamento global (Plano B) e instabilidade climática (Plano C), a fim de elevar o potencial pró-ativo de resiliências nas sociedades.

Made available in DSpace on 2016-08-17T14:52:49Z (GMT). No. of bitstreams: 1 Ismenia Ribeiro de Oliveira.pdf: 2858888 bytes, checksum: 1f919da635a28d5ee15d5832df5d95b1 (MD5) Previous issue date: 2004-02-10 / This work proposes an agent-based pattern system for design of adaptive multi-agent systems using user-modeling techniques. The pattern system is composed of a conceptual, an architectural, and design patterns. The patterns describe recurrent problems and design solutions for user modeling and development of adaptive multi-agent systems. The pattern mining is based on the analysis of: concepts and techniques for development of adaptive and adaptable systems; user modeling techniques; and multi-agent design techniques. An initial validation of the pattern system has been conducted through the development of a case study on the domain of information access, where patterns in the system have been reused on the development of a multi-agent system for personalized information access. / Este trabalho propõe um sistema de padrões baseados em agentes para o projeto de sistemas multiagente adaptativos e adaptáveis que utilizam técnicas de modelagem de usuários. O sistema de padrões é composto por um padrão conceitual, um padrão arquitetural e cinco padrões de projeto. Os problemas, as soluções e os conceitos, comuns no projeto da modelagem de usuários, de sistemas adaptativos e no desenvolvimento de agentes de software, são analisados e descritos na forma de padrões. A metodologia para a extração dos padrões é baseada na análise dos conceitos e técnicas para o desenvolvimento de sistemas adaptativos e adaptáveis, das técnicas de modelagem de usuário e dos conceitos e técnicas para o projeto de sistemas multiagente e agentes de software. Uma avaliação preliminar, do sistema de padrões proposto, é feita por meio de um estudo de caso no domínio do acesso à informação. O estudo consiste em identificar e analisar a reutilização do sistema de padrões no desenvolvimento de aplicações multiagente para o acesso à informação personalizada.

Made available in DSpace on 2014-08-20T14:24:47Z (GMT). No. of bitstreams: 1 Dissertacao Gisele Medina Nunes.pdf: 1856118 bytes, checksum: 605a212410288dd200904246fb033632 (MD5) Previous issue date: 2013-03-08 / In the search for an approach that embraced language learning in a more global and dynamic way, applied linguists such as Larsen-Freeman (1997) and Paiva (2005) resorted to Chaos/Complexity theories as a metaphor to deal with the learning process. In this perspective, we defend that learning a language is a complex adaptive system (CAS), since it has all of its characteristics: it is open, susceptible to external factors and to initial conditions, unpredictable, non-linear, ruled by low level rules and self-organized. Besides, in order to keep functioning, active, a CAS needs the interaction among its constituting elements. In this study, we also worked with the concept of interlanguage, developed through five central processes (SELINKER, 1972) which we understand to have some points in common with the characteristics that constitute the CAS. It is also relevant to ponder the fact that the learning process of a foreign language can be affected by many factors through time, as well as that it does not follow a determined pattern of development, something that could be predicted. Likewise, innate factors should be taken into account in the learning process, since the context in which it takes place also plays an important role in the success of learning a language. Thus, the aim of this research was to investigate, through the writing ability being practiced and shown in a blog, the development of the interlanguage of two students of a teaching course in Languages/English of a public university. The intention was to verify how the interlanguage of the two students behaves during the sixth and seventh semester of English Language when the online activity was developed, analyzing the fluency, accuracy and grammatical complexity of their published texts. Besides, we observed how the collaborative correction of these texts, when the students worked as proofreaders of one another, making linguistic as well as communicative adjustments in the productions using the online tool, can influence their interlanguage, contributing to their improvement in writing in English. / Na busca por uma abordagem que abarcasse a aprendizagem de línguas de uma forma mais global e dinâmica, linguistas aplicados como Larsen-Freeman (1997) e Paiva (2005) recorreram às teorias do Caos/Complexidade como metáfora para tratar desse processo. Nessa perspectiva, defende-se que aprender uma língua é um sistema adaptativo complexo, pois possui todas as suas características: é aberto, sensível a fatores externos e a condições iniciais, imprevisível, não-linear, regido por regras de baixo nível e auto-organizável. Além disso, considera-se também que para um sistema se manter ativo, funcionando, é necessária a interação de suas partes, já que é por meio da interação que os elementos que constituem o sistema são mantidos em constante movimento. Neste estudo também se trabalhou com o conceito de interlíngua, desenvolvida por meio de cinco processos centrais (SELINKER,1972), os quais se entende que possuem pontos de contato com as características que delineiam os sistemas complexos. É também relevante atentar para o fato de que o processo de aprendizagem de um idioma pode ser afetado por diversos fatores ao longo do tempo, além de não seguir um padrão determinado de desenvolvimento que possa ser previsto. Da mesma forma como questões inatas devem ser consideradas na aquisição de uma língua, o meio onde esse processo se dá também desempenha um papel importante para o sucesso na aprendizagem. Assim, o foco da presente pesquisa foi investigar o desenvolvimento da interlíngua em duas alunas de um curso de Licenciatura em Letras Habilitação em Língua Inglesa de uma universidade federal por meio da observação da presença da interlíngua na habilidade escrita trabalhada em um blog. Pretendeu-se verificar, por meio da fluência, precisão e complexidade gramatical dos textos publicados pelas alunas, como sua interlíngua comportou-se durante o sexto e o sétimo semestres da disciplina de Língua Inglesa em que a atividade de produção online nessa língua foi proposta. Além disso, se observou como o trabalho colaborativo mediado pela ferramenta online, em que os colegas agiram como revisores dos comunicativa, pode influenciar a interlíngua dessas alunas, contribuindo para seu aprimoramento na escrita no idioma.

Denna kandidatuppsats undersöker hur user experience kan främjas i ett adaptivt talbaserat system i en fordonskontext. Att undersöka och involvera user experiencei adaptiva talbaserade system är i synnerhet viktigt i denna kontext då den primära sysslan, att köra, bör störas så lite som möjligt. Ett adaptivt system ger möjlighet att effektivisera interaktion och ett talbaserat system låter föraren fokusera på trafiken.Genom att undersöka existerande teori och utföra empiriska undersökningar i form av användartester genom Wizard of Oz-tekniken och intervjuer har vi skapat designoch dialogexempel. Dessa förslag kan främja user experience och på så vis bidra till att den interaktion som sker är effektiv och behaglig. / This bachelor thesis aims to build an understanding how user experience can be promoted in an adaptive speech-based system, which is found in the context of an automobile. Investigating this particular context and system is of importance since the primary task is to drive safely and let the driver focus on driving rather than system. An adaptive system provides the opportunity to make the interaction more effective and a speech-based system lets the driver focus on the traffic environment. By studying relevant theory and by carrying out user tests, with the technique Wizard of Oz, we have drafted design and dialogue recommendations. These recommendations can promote user experience and make the interaction of the system more effective and pleasant. / Safe Speech for Knowledge (SSK)

Abstract Smart adaptive systems provide advanced tools for monitoring, control, diagnostics and management of nonlinear multivariate processes. Data mining with a multitude of methodologies is a good basis for the integration of intelligent systems. Small, specialised systems have a large number of feasible solutions, but highly complex systems require domain expertise and more compact approaches at the basic level. Linguistic equation (LE) approach originating from fuzzy logic is an efficient technique for these problems. This research is focused on the smart adaptive applications, where different intelligent modules are used in a smart way. The nonlinear scaling methodology based on advanced statistical analysis is the corner stone in representing the variable meanings in a compact way to introduce intelligent indices for control and diagnostics. The new constraint handling together with generalised norms and moments facilitates recursive parameter estimation approaches for the adaptive scaling. Well-known linear methodologies are used for the steady state, dynamic and case-based modelling in connection with the cascade and interactive structures in building complex large scale applications. To achieve insight and robustness the parameters are defined separately for the scaling and the interactions. The LE based intelligent analysers are useful in the multilevel LE control and diagnostics: the LE control is enhanced with the intelligent analysers, adaptive and model-based modules and high level control. The operating area is extended with the predefined adaptation and specific events activate appropriate control actions. The condition, stress and trend indices are used for the detection of operating conditions. The same overall structure is extended to the scheduling and managerial decision support. The linguistic representation becomes increasingly important when the human interaction is essential. The new scaling approach is used in control and diagnostic applications and discussed in connection with previous multivariate modelling cases. The LE based intelligent analysers are the key modules of the system integration, which produces hybrid systems: fuzzy systems move gradually to higher levels, neural networks and evolutionary computing are used for tuning. The overall system is reinforced with advanced statistical analysis, signal processing, feature extraction, classification and mechanistic modelling. / Tiivistelmä Viisaat mukautuvat järjestelmät sisältävät kehittyneitä työkaluja epälineaaristen monimuuttujaisten prosessien valvontaan, säätöön, diagnostiikkaan ja johtamiseen. Laajaan menetelmäpohjaan perustuva tiedonrikastus on pohjana älykkäiden järjestelmien yhdistämiselle. Pienille erikoistuneille järjestelmille on monia toteutettavissa olevia ratkaisuja, mutta erittäin monimutkaiset järjestelmät vaativat alan asiantuntemusta ja kompakteja lähestymistapoja perustasolla. Sumeaan logiikkaan pohjautuva lingvististen yhtälöiden (linguistic equation, LE) menetelmä on tehokas ratkaisu näissä ongelma-alueissa. Tämä tutkimus kohdistuu viisaisiin mukautuviin sovelluksiin, jossa useita älykkäitä moduuleja käytetään yhdessä viisaalla tavalla. Kehittyneeseen tilastolliseen analyysiin perustuva epälineaarinen skaalausmenetelmä muodostaa ratkaisun kulmakiven: muuttujien merkitykset soveltuvat säädössä ja diagnostiikassa käytettävien älykkäiden indeksien kehittämiseen. Uudet rajoituksien käsittelymenetelmät yhdessä yleistettyjen normien ja momenttien kanssa mahdollistavat rekursiivisen parametriestimoinnin olosuhteisiin mukautuvassa skaalauksessa. Tunnettuja lineaarisia menetelmiä käytetään staattisessa, dynaamisessa ja tapauspohjaisessa mallintamisessa, jossa kaskadi- ja vuorovaikutusrakenteet laajentavat mallit tarvittaessa monimutkaisiin sovelluksiin. Prosessituntemuksen ja järjestelmien robustisuuden varmistamiseksi parametrit määritellään erikseen skaalausta ja vuorovaikutuksia varten. LE-pohjaiset älykkäät analysaattorit ovat hyödyllisiä monitasoisessa säädössä ja diagnostiikassa: LE-säätöä parannetaan älykkäiden analysaattorien, adaptiivisten ja mallipohjaisten moduulien sekä ylemmän tason säädön avulla. Käyttöaluetta laajennetaan ennalta määrätyllä adaptoinnilla sekä tiettyjen tapahtumien aktivoimilla erityisillä säätötoimenpiteillä. Kunto-, rasitus- ja trendi-indeksejä käytetään olosuhteiden tunnistamiseen. Sama rakenne laajennetaan tuotannon ajoitukseen ja päätöksenteontukeen, jossa inhimillisen vuorovaikutuksen käsittely tekee lingvistisen esityksen yhä tärkeämmäksi. Uutta skaalausmenetelmää tarkastellaan säätö- ja diagnostiikkasovelluksissa sekä vertaillaan lyhyesti sen käyttömahdollisuuksia aikaisemmin toteutetuissa monimuuttujamalleissa. LE-pohjaiset älykkäät analysaattorit ovat keskeisiä integroitaessa moduuleja hybridiratkaisuiksi: sumeat järjestelmät siirtyvät vähitellen ylemmille tasoille ja neuro- ja evoluutiolaskennassa keskitytään järjestelmien viritykseen. Kokonaisjärjestelmää vahvistetaan kehittyneellä tilastollisella analyysilla, signaalinkäsittelyllä, piirteiden erottamisella, luokittelulla ja mekanistisella mallintamisella.

The work presented in this dissertation investigates techniques for robot Learning from Demonstration (LFD). LFD is a well established approach where the robot is to learn from a set of demonstrations. The dissertation focuses on LFD where a human teacher demonstrates a behavior by controlling the robot via teleoperation. After demonstration, the robot should be able to reproduce the demonstrated behavior under varying conditions. In particular, the dissertation investigates techniques where previous behavioral knowledge is used as bias for generalization of demonstrations. The primary contribution of this work is the development and evaluation of a semi-reactive approach to LFD called Predictive Sequence Learning (PSL). PSL has many interesting properties applied as a learning algorithm for robots. Few assumptions are introduced and little task-specific configuration is needed. PSL can be seen as a variable-order Markov model that progressively builds up the ability to predict or simulate future sensory-motor events, given a history of past events. The knowledge base generated during learning can be used to control the robot, such that the demonstrated behavior is reproduced. The same knowledge base can also be used to recognize an on-going behavior by comparing predicted sensor states with actual observations. Behavior recognition is an important part of LFD, both as a way to communicate with the human user and as a technique that allows the robot to use previous knowledge as parts of new, more complex, controllers. In addition to the work on PSL, this dissertation provides a broad discussion on representation, recognition, and learning of robot behavior. LFD-related concepts such as demonstration, repetition, goal, and behavior are defined and analyzed, with focus on how bias is introduced by the use of behavior primitives. This analysis results in a formalism where LFD is described as transitions between information spaces. Assuming that the behavior recognition problem is partly solved, ways to deal with remaining ambiguities in the interpretation of a demonstration are proposed. The evaluation of PSL shows that the algorithm can efficiently learn and reproduce simple behaviors. The algorithm is able to generalize to previously unseen situations while maintaining the reactive properties of the system. As the complexity of the demonstrated behavior increases, knowledge of one part of the behavior sometimes interferes with knowledge of another parts. As a result, different situations with similar sensory-motor interactions are sometimes confused and the robot fails to reproduce the behavior. One way to handle these issues is to introduce a context layer that can support PSL by providing bias for predictions. Parts of the knowledge base that appear to fit the present context are highlighted, while other parts are inhibited. Which context should be active is continually re-evaluated using behavior recognition. This technique takes inspiration from several neurocomputational models that describe parts of the human brain as a hierarchical prediction system. With behavior recognition active, continually selecting the most suitable context for the present situation, the problem of knowledge interference is significantly reduced and the robot can successfully reproduce also more complex behaviors.

In this thesis, research efforts are dedicated towards the development of practical adaptable techniques to be used in Smart Homes and Buildings, with the aim to improve energy management and conservation, while enhancing the learning capabilities of Programmable Communicating Thermostats (PCT) – “transforming” them into smart adaptable devices, i.e., “Smart Thermostats”. An Adaptable Hybrid Intelligent System utilizing Wireless Sensor Network (WSN) and Artificial Intelligence (AI) techniques is presented, based on which, a novel Adaptive Learning System (ALS) model utilizing WSN, a rule-based system and Adaptive Resonance Theory (ART) concepts is proposed. The main goal of the ALS is to adapt to the occupant’s pattern and/or schedule changes by providing comfort, while not ignoring the energy conservation aspect. The proposed ALS analytical model is a technique which enables PCTs to learn and adapt to user input pattern changes and/or other parameters of interest. A new algorithm for finding the global maximum in a predefined interval within a two dimensional space is proposed. The proposed algorithm is a synergy of reward/punish concepts from the reinforcement learning (RL) and agent-based technique, for use in small-scale embedded systems with limited memory and/or processing power, such as the wireless sensor/actuator nodes. An application is implemented to observe the algorithm at work and to demonstrate its main features. It was observed that the “RL and Agent-based Search”, versus the “RL only” technique, yielded better performance results with respect to the number of iterations and function evaluations needed to find the global maximum. Furthermore, a “House Simulator” is developed as a tool to simulate house heating/cooling systems and to assist in the practical implementation of the ALS model under different scenarios. The main building blocks of the simulator are the “House Simulator”, the “Smart Thermostat”, and a placeholder for the “Adaptive Learning Models”. As a result, a novel adaptive learning algorithm, “Observe, Learn and Adapt” (OLA) is proposed and demonstrated, reflecting the main features of the ALS model. Its evaluation is achieved with the aid of the “House Simulator”. OLA, with the use of sensors and the application of the ALS model learning technique, captures the essence of an actual PCT reflecting a smart and adaptable device. The experimental performance results indicate adaptability and potential energy savings of the single in comparison to the zone controlled scenarios with the OLA capabilities being enabled.

In the domain of dynamically evolving distributed systems composed of autonomous and (self-) adaptive components, the task of systematically managing the design complexity of their communication and composition is a pressing issue. This stems from the dynamic nature of such systems, where components and their bindings may appear and disappear without anticipation. As one way of addressing the challenge, [15] proposes a new component model (called DEECo), which features separation of concerns via a mechanism of dynamic implicit bindings with implicit communication. This leads to introduction of autonomous components and their dynamic interaction groups (called ensembles). The goal of the thesis is to reify the DEECo concepts and paradigms in a Java-based implementation connected to a simulation of a virtual world. As such, the thesis should provide a platform for experimentation with DEECo-based applications.

Future applications will run in a highly heterogeneous and dynamic execution environment that forces them to adapt their behavior and offered functionality depending on the user's or the system's current situation. Since application components in such heterogeneous multi-device systems will be distributed over multiple interconnected devices and cooperate to achieve a common goal, a coordinated adaptation is required to ensure a consistent system behavior. In this paper we present a decentralized adaptation middleware to adapt a distributed software system. Our approach supports the reliable execution of multiple adaptation operations that depend on each other and are performed transactionally even in unsteady environments coined by message loss or node failures. We implemented our approach in a search-and-rescue robot scenario to show its feasibility and conduct first performance evaluations.