Optimising market share and profit margin:SMDP-based tariff pricing under the smart grid paradigm

Smart grid technologies have given rise to a liberalised and decentralised electricity market, enabling energy providers and retailers to have a better understanding of the demand side and its response to pricing signals. This paper puts forward a reinforcement-learning-powered tool aiding an electricity retailer to define the tariff prices it offers, in a bid to optimise its retail strategy. In a competitive market, an energy retailer aims to simultaneously increase the number of contracted customers and its profit margin. We have abstracted the problem of deciding on a tariff price as faced by a retailer, as a semi-Markov decision problem (SMDP). A hierarchical reinforcement learning approach, MaxQ value function decomposition, is applied to solve the SMDP through interactions with the market. To evaluate our trading strategy, we developed a retailer agent (termed AstonTAC) that uses the proposed SMDP framework to act in an open multi-agent simulation environment, the Power Trading Agent Competition (Power TAC). An evaluation and analysis of the 2013 Power TAC finals show that AstonTAC successfully selects sell prices that attract as many customers as necessary to maximise the profit margin. Moreover, during the competition, AstonTAC was the only retailer agent performing well across all retail market settings.

Intelligent Car Parking Locator Service

This paper presents an InfoStation-based multi-agent system facilitating a Car Parking Locator service provision within a University Campus. The system network architecture is outlined, illustrating its functioning during the service provision. A detailed description of the Car Parking Locator service is given and the system entities’ interaction is described. System implementation approaches are also considered.

InfoStation-based Adaptable Provision of m Learning Services: Main Scenarios

This paper presents an adaptable InfoStation-based multi-agent system facilitating the mobile eLearning (mLearning) service provision within a University Campus. A horizontal view of the network architecture is presented. Main communications scenarios are considered by describing the detailed interaction of the system entities involved in the mLearning service provision. The mTest service is explored as a practical example. System implementation approaches are also considered.

A mixture density network approach to modelling and exploiting uncertainty in nonlinear control problems

In nonlinear and stochastic control problems, learning an efficient feed-forward controller is not amenable to conventional neurocontrol methods. For these approaches, estimating and then incorporating uncertainty in the controller and feed-forward models can produce more robust control results. Here, we introduce a novel inversion-based neurocontroller for solving control problems involving uncertain nonlinear systems which could also compensate for multi-valued systems. The approach uses recent developments in neural networks, especially in the context of modelling statistical distributions, which are applied to forward and inverse plant models. Provided that certain conditions are met, an estimate of the intrinsic uncertainty for the outputs of neural networks can be obtained using the statistical properties of networks. More generally, multicomponent distributions can be modelled by the mixture density network. Based on importance sampling from these distributions a novel robust inverse control approach is obtained. This importance sampling provides a structured and principled approach to constrain the complexity of the search space for the ideal control law. The developed methodology circumvents the dynamic programming problem by using the predicted neural network uncertainty to localise the possible control solutions to consider. A nonlinear multi-variable system with different delays between the input-output pairs is used to demonstrate the successful application of the developed control algorithm. The proposed method is suitable for redundant control systems and allows us to model strongly non-Gaussian distributions of control signal as well as processes with hysteresis. © 2004 Elsevier Ltd. All rights reserved.

BOWL:augmenting the semantic web with beliefs

As the Semantic Web is an open, complex and constantly evolving medium, it is the norm, but not exception that information at different sites is incomplete or inconsistent. This poses challenges for the engineering and development of agent systems on the Semantic Web, since autonomous software agents need to understand, process and aggregate this information. Ontology language OWL provides core language constructs to semantically markup resources on the Semantic Web, on which software agents interact and cooperate to accomplish complex tasks. However, as OWL was designed on top of (a subset of) classic predicate logic, it lacks the ability to reason about inconsistent or incomplete information. Belief-augmented Frames (BAF) is a frame-based logic system that associates with each frame a supporting and a refuting belief value. In this paper, we propose a new ontology language Belief-augmented OWL (BOWL) by integrating OWL DL and BAF to incorporate the notion of confidence. BOWL is paraconsistent, hence it can perform useful reasoning services in the presence of inconsistencies and incompleteness. We define the abstract syntax and semantics of BOWL by extending those of OWL. We have proposed reasoning algorithms for various reasoning tasks in the BOWL framework and we have implemented the algorithms using the constraint logic programming framework. One example in the sensor fusion domain is presented to demonstrate the application of BOWL.

A New Class of Processes for Formalizing and Generalizing Individual-Based Models: The Semi-Semi-Markov Processes

2000 Mathematics Subject Classification: 60K15, 60K20, 60G20,60J75, 60J80, 60J85, 60-08, 90B15.

Maximising the efficiency of logistics operations with radio frequency identification technology

Radio frequency identification (RFID) technology has gained increasing popularity in businesses to improve operational efficiency and maximise costs saving. However, there is a gap in the literature exploring the enhanced use of RFID to substantially add values to the supply chain operations, especially beyond what the RFID vendors could offer. This paper presents a multi-agent system, incorporating RFID technology, aimed at fulfilling the gap. The system is developed to model supply chain activities (in particular, logistics operations) and is comprised of autonomous and intelligent agents representing the key entities in the supply chain. With the advanced characteristics of RFID incorporated, the agent system examines ways logistics operations (i.e. distribution network) particular) can be efficiently reconfigured and optimised in response to dynamic changes in the market, production and at any stage in the supply chain. © 2012 IEEE.

Exotic meson decay widths using lattice quantum chromodynamics

Most experiments in particle physics are scattering experiments, the analysis of which leads to masses, scattering phases, decay widths and other properties of one or multi-particle systems. Until the advent of Lattice Quantum Chromodynamics (LQCD) it was difficult to compare experimental results on low energy hadron-hadron scattering processes to the predictions of QCD, the current theory of strong interactions. The reason being, at low energies the QCD coupling constant becomes large and the perturbation expansion for scattering; amplitudes does not converge. To overcome this, one puts the theory onto a lattice, imposes a momentum cutoff, and computes the integral numerically. For particle masses, predictions of LQCD agree with experiment, but the area of decay widths is largely unexplored. ^ LQCD provides ab initio access to unusual hadrons like exotic mesons that are predicted to contain real gluonic structure. To study decays of these type resonances the energy spectra of a two-particle decay state in a finite volume of dimension L can be related to the associated scattering phase shift δ(k) at momentum k through exact formulae derived by Lüscher. Because the spectra can be computed using numerical Monte Carlo techniques, the scattering phases can thus be determined using Lüscher's formulae, and the corresponding decay widths can be found by fitting Breit-Wigner functions. ^ Results of such a decay width calculation for an exotic hybrid( h) meson (JPC = 1-+) are presented for the decay channel h → πa 1. This calculation employed Lüscher's formulae and an approximation of LQCD called the quenched approximation. Energy spectra for the h and πa1 systems were extracted using eigenvalues of a correlation matrix, and the corresponding scattering phase shifts were determined for a discrete set of πa1 momenta. Although the number of phase shift data points was sparse, fits to a Breit-Wigner model were made, resulting in a decay width of about 60 MeV. ^

Utilizando comitês de classificadores para predição de rendimento escolar

Educational Data Mining is an application domain in artificial intelligence area that has been extensively explored nowadays. Technological advances and in particular, the increasing use of virtual learning environments have allowed the generation of considerable amounts of data to be investigated. Among the activities to be treated in this context exists the prediction of school performance of the students, which can be accomplished through the use of machine learning techniques. Such techniques may be used for student’s classification in predefined labels. One of the strategies to apply these techniques consists in their combination to design multi-classifier systems, which efficiency can be proven by results achieved in other studies conducted in several areas, such as medicine, commerce and biometrics. The data used in the experiments were obtained from the interactions between students in one of the most used virtual learning environments called Moodle. In this context, this paper presents the results of several experiments that include the use of specific multi-classifier systems systems, called ensembles, aiming to reach better results in school performance prediction that is, searching for highest accuracy percentage in the student’s classification. Therefore, this paper presents a significant exploration of educational data and it shows analyzes of relevant results about these experiments.

Context-Aware Personalised Service Delivery

Postprint

Using phase-space localized basis functions to obtain vibrational energies of molecules

For decades scientists have attempted to use ideas of classical mechanics to choose basis functions for calculating spectra. The hope is that a classically-motivated basis set will be small because it covers only the dynamically important part of phase space. One popular idea is to use phase space localized (PSL) basis functions. This thesis improves on previous efforts to use PSL functions and examines the usefulness of these improvements. Because the overlap matrix, in the matrix eigenvalue problem obtained by using PSL functions with the variational method, is not an identity, it is costly to use iterative methods to solve the matrix eigenvalue problem. We show that it is possible to circumvent the orthogonality (overlap) problem and use iterative eigensolvers. We also present an altered method of calculating the matrix elements that improves the performance of the PSL basis functions, and also a new method which more efficiently chooses which PSL functions to include. These improvements are applied to a variety of single well molecules. We conclude that for single minimum molecules, the PSL functions are inferior to other basis functions. However, the ideas developed here can be applied to other types of basis functions, and PSL functions may be useful for multi-well systems.

Human behaviour recognition in data-scarce domains

This paper presents the novel theory for performing multi-agent activity recognition without requiring large training corpora. The reduced need for data means that robust probabilistic recognition can be performed within domains where annotated datasets are traditionally unavailable. Complex human activities are composed from sequences of underlying primitive activities. We do not assume that the exact temporal ordering of primitives is necessary, so can represent complex activity using an unordered bag. Our three-tier architecture comprises low-level video tracking, event analysis and high-level inference. High-level inference is performed using a new, cascading extension of the Rao–Blackwellised Particle Filter. Simulated annealing is used to identify pairs of agents involved in multi-agent activity. We validate our framework using the benchmarked PETS 2006 video surveillance dataset and our own sequences, and achieve a mean recognition F-Score of 0.82. Our approach achieves a mean improvement of 17% over a Hidden Markov Model baseline.

A Collaborative Multiagent Framework based on Online Risk-Aware Planning and Decision-Making

Planning is an essential process in teams of multiple agents pursuing a common goal. When the effects of actions undertaken by agents are uncertain, evaluating the potential risk of such actions alongside their utility might lead to more rational decisions upon planning. This challenge has been recently tackled for single agent settings, yet domains with multiple agents that present diverse viewpoints towards risk still necessitate comprehensive decision making mechanisms that balance the utility and risk of actions. In this work, we propose a novel collaborative multi-agent planning framework that integrates (i) a team-level online planner under uncertainty that extends the classical UCT approximate algorithm, and (ii) a preference modeling and multicriteria group decision making approach that allows agents to find accepted and rational solutions for planning problems, predicated on the attitude each agent adopts towards risk. When utilised in risk-pervaded scenarios, the proposed framework can reduce the cost of reaching the common goal sought and increase effectiveness, before making collective decisions by appropriately balancing risk and utility of actions. 

Uma abordagem evolucionária para o planejamento online de rotas para múltiplos robôs móveis em ambiente multiagente

O problema de planejamento de rotas de robôs móveis consiste em determinar a melhor rota para um robô, em um ambiente estático e/ou dinâmico, que seja capaz de deslocá-lo de um ponto inicial até e um ponto final, também em conhecido como estado objetivo. O presente trabalho emprega o uso de uma abordagem baseada em Algoritmos Genéticos para o planejamento de rotas de múltiplos robôs em um ambiente complexo composto por obstáculos fixos e obstáculos moveis. Através da implementação do modelo no software do NetLogo, uma ferramenta utilizada em simulações de aplicações multiagentes, possibilitou-se a modelagem de robôs e obstáculos presentes no ambiente como agentes interativos, viabilizando assim o desenvolvimento de processos de detecção e desvio de obstáculos. A abordagem empregada busca pela melhor rota para robôs e apresenta um modelo composto pelos operadores básicos de reprodução e mutação, acrescido de um novo operador duplo de refinamento capaz de aperfeiçoar as melhores soluções encontradas através da eliminação de movimentos inúteis. Além disso, o calculo da rota de cada robô adota um método de geração de subtrechos, ou seja, não calcula apenas uma unica rota que conecta os pontos inicial e final do cenário, mas sim várias pequenas subrotas que conectadas formam um caminho único capaz de levar o robô ao estado objetivo. Neste trabalho foram desenvolvidos dois cenários, para avaliação da sua escalabilidade: o primeiro consiste em um cenário simples composto apenas por um robô, um obstáculo movel e alguns obstáculos fixos; já o segundo, apresenta um cenário mais robusto, mais amplo, composto por múltiplos robôs e diversos obstáculos fixos e moveis. Ao final, testes de desempenho comparativos foram efetuados entre a abordagem baseada em Algoritmos Genéticos e o Algoritmo A*. Como critério de comparação foi utilizado o tamanho das rotas obtidas nas vinte simulações executadas em cada abordagem. A analise dos resultados foi especificada através do Teste t de Student.

Fighting Evasive Malware with DVasion

Malware is a foundational component of cyber crime that enables an attacker to modify the normal operation of a computer or access sensitive, digital information. Despite the extensive research performed to identify such programs, existing schemes fail to detect evasive malware, an increasingly popular class of malware that can alter its behavior at run-time, making it difficult to detect using today’s state of the art malware analysis systems. In this thesis, we present DVasion, a comprehensive strategy that exposes such evasive behavior through a multi-execution technique. DVasion successfully detects behavior that would have been missed by traditional, single-execution approaches, while addressing the limitations of previously proposed multi-execution systems. We demonstrate the accuracy of our system through strong parallels with existing work on evasive malware, as well as uncover the hidden behavior within 167 of 1,000 samples.