Efforts towards a low carbon economy are challenging the electricity industry. On the supply-side, centralised carbon-intensive power plants are set to gradually decrease their contribution to the generation mix, whilst distributed renewable generation is to successively increase its share. On the demand-side, electricity use is expected to increase in the future due to the electrification of heating and transport. Moreover, the demand-side is to become more active allowing end-users to invest in generation and storage technologies, such as solar photovoltaics (PV) and home batteries. As a result, some network reinforcements might be needed and instrumentation at the users’ end is to be required, such as controllers and home energy management systems (HEMS). The electricity grid must balance supply and demand at all times in order to maintain technical constraints of frequency, voltage, and current; and this will become more challenging as a result of this transition. Failure to meet these constraints compromises the service and could damage the power grid assets and end-users’ appliances. Balancing generation, although responsive, is carbon-intensive and associated with inefficient asset utilisation, as these generators are mostly used during peak hours and sit idle the rest of the time. Furthermore, energy storage is a potential solution to assist the balancing problem in the presence of non-dispatchable low-carbon generators; however, it is substantially expensive to store energy in large amounts. Therefore, demand response (DR) has been envisioned as a complementary solution to increase the system’s resilience to weather-dependent, stochastic, and intermittent generation along with variable and temperature-correlated electric load. In the domestic setting, operational flexibility of some appliances, such as heaters and electric cars, can be coordinated amongst several households so as to help balance supply and demand, and reduce the need of balancing generators. Against this background, the electricity supply system requires new organisational paradigms that integrate DR effectively. Although some dynamic pricing schemes have been proposed to guide DR, such as time of use (ToU) and real-time pricing (RTP), it is still unclear how to control oscillatory massive responses (e.g., large fleet of electric cars simultaneously responding to a favourable price). Hence, this thesis proposes an alternative approach in which households proactively submit DR offers that express their preferences to their respective retailer in exchange for a discount. This research develops a computational model of domestic electricity use, and simulates appliances with operational flexibility in order to evaluate the effects and benefits of DR for both retailers and households. It provides a representation for this flexibility so that it can be integrated into specific DR offers. Retailers and households are modelled as computational agents. Furthermore, two market-based mechanisms are proposed to determine the allocation of DR offers. More specifically, a one-sided Vickrey-Clarke-Groves (VCG)-based mechanism and penalty schemes were designed for electricity retailers to coordinate their customers’ DR efforts so as to ameliorate the imbalance of their trading schedules. Similarly, a two-sided McAfee-based mechanism was designed to integrate DR offers into a multi-retailer setting in order to reduce zonal imbalances. A suitable method was developed to construct DR block offers that could be traded amongst retailers. Both mechanisms are dominant-strategy incentive-compatible and trade off a small amount of economic efficiency in order to maintain individual rationality, truthful reporting, weak budget balance and tractable computation. Moreover, privacy preserving is achieved by including computational agents from the independent system operator (ISO) as intermediaries between each retailer and its domestic customers, and amongst retailers. The theoretical properties of these mechanisms were proved using worst-case analysis, and their economic effects were evaluated in simulations based on data from a survey of UK household electricity use. In addition, forecasting methods were assessed on the end-users’ side in order to make better DR offers and avoid penalties. The results show that, under reasonable assumptions, the proposed coordination mechanisms achieve significant savings for both end-users and retailers, as they reduce the required amount of expensive balancing generation.

Ao tentar agendar uma atividade que dependa da presença de outras pessoas, geralmente acabamos desperdiçando tempo precioso avaliando os possíveis horários e verificando se os mesmos são aceitos por todos envolvidos. Embora a modelagem e a resolução do problema de agendamento multiagente pareçam estar completamente entendidas e ainda diversos algoritmos possam ser encontrados na literatura, uma questão ainda existe: Como definir horários compatíveis para uma atividade compartilhada sem que os usuários tenham que manualmente escolher horários livres de seus calendários até que todos envolvidos aceitem um horário. A principal contribuição é um algoritmo chamado Descobridor Asíncrono de Horários (ATF) baseado no Rastreamento Asíncrono (ABT) que permite que aplicações encontrem horários compatíveis para atividades compartilhadas requerendo mínima intervenção manual dos usuários. Esta dissertação revisita o Problema Temporal Simples (STP) e a sua versão multiagente (MaSTP), demonstra como eles podem ser utilizados para resolver o problema de agentamentos e ao final apresenta o ATF, a avaliação experimental e a análise de complexidade. / In order to schedule an activity that depends on other people, we very often end up wasting precious time trying to find compatible times and evaluating if they are accepted by all involved. Even though modeling and solving multiagent scheduling problems seem completely understood and several algorithms can be found in the literature, one limitation still stands up: How to find a compatible time slot for an activity shared by many users without requiring the users themselves to spend time going through their calendar and choosing time slots until everybody agrees. The main contribution of this work is an algorithm called Asynchronous Time Finder (ATF) based on the Asynchronous Backtracking (ABT) that enables applications to find compatible times when scheduling shared activities among several users while requiring minimal user interaction. This dissertation starts by revisiting the Simple Temporal Problem (STP) and its multiagent version (MaSTP), it then shows how they can be used to solve the problem of managing agendas and then finally it presents the ATF giving an experimental evaluation and the analysis of its complexity.

Ao tentar agendar uma atividade que dependa da presença de outras pessoas, geralmente acabamos desperdiçando tempo precioso avaliando os possíveis horários e verificando se os mesmos são aceitos por todos envolvidos. Embora a modelagem e a resolução do problema de agendamento multiagente pareçam estar completamente entendidas e ainda diversos algoritmos possam ser encontrados na literatura, uma questão ainda existe: Como definir horários compatíveis para uma atividade compartilhada sem que os usuários tenham que manualmente escolher horários livres de seus calendários até que todos envolvidos aceitem um horário. A principal contribuição é um algoritmo chamado Descobridor Asíncrono de Horários (ATF) baseado no Rastreamento Asíncrono (ABT) que permite que aplicações encontrem horários compatíveis para atividades compartilhadas requerendo mínima intervenção manual dos usuários. Esta dissertação revisita o Problema Temporal Simples (STP) e a sua versão multiagente (MaSTP), demonstra como eles podem ser utilizados para resolver o problema de agentamentos e ao final apresenta o ATF, a avaliação experimental e a análise de complexidade. / In order to schedule an activity that depends on other people, we very often end up wasting precious time trying to find compatible times and evaluating if they are accepted by all involved. Even though modeling and solving multiagent scheduling problems seem completely understood and several algorithms can be found in the literature, one limitation still stands up: How to find a compatible time slot for an activity shared by many users without requiring the users themselves to spend time going through their calendar and choosing time slots until everybody agrees. The main contribution of this work is an algorithm called Asynchronous Time Finder (ATF) based on the Asynchronous Backtracking (ABT) that enables applications to find compatible times when scheduling shared activities among several users while requiring minimal user interaction. This dissertation starts by revisiting the Simple Temporal Problem (STP) and its multiagent version (MaSTP), it then shows how they can be used to solve the problem of managing agendas and then finally it presents the ATF giving an experimental evaluation and the analysis of its complexity.

Ao tentar agendar uma atividade que dependa da presença de outras pessoas, geralmente acabamos desperdiçando tempo precioso avaliando os possíveis horários e verificando se os mesmos são aceitos por todos envolvidos. Embora a modelagem e a resolução do problema de agendamento multiagente pareçam estar completamente entendidas e ainda diversos algoritmos possam ser encontrados na literatura, uma questão ainda existe: Como definir horários compatíveis para uma atividade compartilhada sem que os usuários tenham que manualmente escolher horários livres de seus calendários até que todos envolvidos aceitem um horário. A principal contribuição é um algoritmo chamado Descobridor Asíncrono de Horários (ATF) baseado no Rastreamento Asíncrono (ABT) que permite que aplicações encontrem horários compatíveis para atividades compartilhadas requerendo mínima intervenção manual dos usuários. Esta dissertação revisita o Problema Temporal Simples (STP) e a sua versão multiagente (MaSTP), demonstra como eles podem ser utilizados para resolver o problema de agentamentos e ao final apresenta o ATF, a avaliação experimental e a análise de complexidade. / In order to schedule an activity that depends on other people, we very often end up wasting precious time trying to find compatible times and evaluating if they are accepted by all involved. Even though modeling and solving multiagent scheduling problems seem completely understood and several algorithms can be found in the literature, one limitation still stands up: How to find a compatible time slot for an activity shared by many users without requiring the users themselves to spend time going through their calendar and choosing time slots until everybody agrees. The main contribution of this work is an algorithm called Asynchronous Time Finder (ATF) based on the Asynchronous Backtracking (ABT) that enables applications to find compatible times when scheduling shared activities among several users while requiring minimal user interaction. This dissertation starts by revisiting the Simple Temporal Problem (STP) and its multiagent version (MaSTP), it then shows how they can be used to solve the problem of managing agendas and then finally it presents the ATF giving an experimental evaluation and the analysis of its complexity.

This thesis is concerned with the ad hoc coordination problem. Therein, the goal is to design an autonomous agent which can achieve high flexibility and efficiency in a multiagent system that admits no prior coordination between the designed agent and the other agents. Flexibility describes the agent’s ability to solve its task with a variety of other agents in the system; efficiency is the relation between the agent’s payoffs and time needed to solve the task; and no prior coordination means that the agent does not a priori know how the other agents behave. This problem is relevant for a number of practical applications, including human-machine interaction tasks, such as adaptive user interfaces, robotic elderly care, and automated trading agents. Motivated by this problem, the central idea studied in this thesis is to utilise a set of policies, or types, to characterise the behaviour of other agents. Specifically, the idea is to reduce the complexity of the interaction problem by assuming that the other agents draw their latent type from some known or hypothesised space of types, and that the assignment of types is governed by an unknown distribution. Based on the current interaction history, we can form posterior beliefs about the relative likelihood of types. These beliefs, combined with the future predictions of the types, can then be used in a planning procedure to compute optimal responses. The aim of this thesis is to study the potential and limitations of this idea in the context of ad hoc coordination. We formulate the ad hoc coordination problem using a game-theoretic model called the stochastic Bayesian game. Based on this model, we derive a canonical algorithmic description of the idea outlined above, called Harsanyi-Bellman Ad Hoc Coordination (HBA). The practical potential of HBA is demonstrated in two case studies, including a human-machine experiment and a simulated logistics domain. We formulate basic ways to incorporate evidence (i.e. observed actions) into posterior beliefs and analyse the conditions under which the posterior beliefs converge to the true distribution of types. Furthermore, we study the impact of prior beliefs over types (that is, before any actions are observed) on the long-term performance of HBA, and show empirically that automatic methods can compute prior beliefs with consistent performance effects. For hypothesised (i.e. “guessed”) type spaces, we analyse the relations between hypothesised and true type spaces under which HBA is still guaranteed to solve its task, despite inaccuracies in hypothesised types. Finally, we show how HBA can perform an automatic statistical analysis to decide whether to reject its behavioural hypothesis, i.e. the combination of posterior beliefs and types.

Ph.D. (Computer Science) / Thanks to several advancements in communication technologies, the world today is a highly connected society promoting business transformations that highlight improved efficiency [1]. Unfortunately, systems developed for an increasingly connected world are also subject to increases in change, complexity and risk – the same connectedness that makes lives easier also signifies that any negative influences can be more difficult to handle and contain [2]. Multi-agent systems have been touted as ideal solutions to realising the required complexities across wide and varied problem domains that range from manufacturing [3] to eco-system management [4] to construction [5]. In an increasingly connected world, complex problems may require that various multi-agent systems work together in order to accomplish larger, overarching objectives. A fraud detection system, for example, could comprise a number of multi-agent systems, each designated to fulfil a very specific and important fraud detection task. The success of the fraud detection system will then depend on each of the various multi-agent systems’ abilities to achieve allocated goals and thus, contribute towards efforts to detect fraud accurately. Depending on factors that include objective and environment type, fraud detection tasks may entail working with numerous disparate systems [6] – it is possible that agent designs that are different from the rest of the fraud detection system must be implemented.Such inconsistency between multi-agent systems could potentially lead to conflicting goals, thereby jeopardising the resolution of the fraud detection system’s overall objectives. A further complication that may arise is the continuously changing financial services landscape – fraud detection systems must not only contend with the creativity of fraudsters, but should also be acutely aware of when day-to-day processes have changed due to recent innovations or technological advancements in the domain. Existing fraud detection methodologies may therefore need to be updated frequently in order to remain sufficiently informed of current developments. An agent-based fraud detection model was thus developed to assist anti-fraud professionals in the classification of day-to-day financial transactions. The proposed model comprises a number of multi-agent systems, each incorporated to add a particular aspect of the criminal justice process in investigating incidences of potential crime. By having agents emulate the various tasks that are involved in dealing with a crime, it is anticipated that the resulting fraud detection system will be able to achieve similar successes from applying the same procedure. In order to successfully develop the fraud detection model, an architecture for implementing a collaborative community of multi-agent subsystems for a dynamic environment was also developed. The architecture is intended to allow each multi-agent subsystem member to adapt to changes in the environment while ensuring that teamwork links are maintained amongst the different subsystems.

M.Sc. (Computer Science) / The purpose of the research presented in this dissertation is to explore the use of a distributed multi-agent system in a general purpose crowd simulation model. Crowd simulation is becoming an increasingly important tool for analysing new construction projects, as it enables safety and performance evaluations to be performed on architectural plans before the buildings have been constructed. Crowd simulation is a challenging problem, as it requires the simulation of complex interactions of people within a crowd. The dissertation investigates existing models of crowd simulation and identifies three primary sub-tasks of crowd simulation: deliberation, path planning and collision-avoiding movement. Deliberation is the process of determining which goal an agent will attempt to satisfy next. Path planning is the process of finding a collision-free path from an agent‟s current location towards its goal. Collision-avoiding movement deals with moving an agent along its calculated path while avoiding collisions with other agents. A multi-agent crowd simulation model, DiMACS, is proposed as a means of addressing the problem of crowd simulation. Multi-agent technology provides an effective solution for representing individuals within a crowd; each member of a crowd can be represented as an intelligent agent. Intelligent agents are capable of maintaining their own internal state and deciding on a course of action based on that internal state. DiMACS is capable of producing realistic simulations while making use of distributed and parallel processing to improve its performance. In addition, the model is highly customisable. The dissertation also presents a user-friendly method for configuring agents within a simulation that abstracts the complexity of agent behaviour away from a user so as to increase the accessibility of configuring the proposed model. In addition, an application programming interface is provided that enables developers to extend the model to simulate additional agent behaviours. The research shows how distributed and parallel processing may be used to improve the performance of an agent-based crowd simulation without compromising the accuracy of the simulation.

M.Sc. (Computer Science) / Artificial intelligence strives towards providing an autonomous mechanism by which the environments of humans may be affected beneficially. The steps taken towards this goal have been to create individual computer programs that solve small problems; however, larger world problems need to be addressed. Intelligence in computer systems cannot be seen as a single algorithm which solves all problems, but rather a set of distinctive algorithms which may be combined uniquely in order to achieve a particular goal. One field of application for artificial intelligence in service to humanity is robotics. Autonomous robotic entities are becoming more commonplace in society, making their behaviour an important topic of study. Machines capable of performing various activities in service of the human race are fundamentally important if they are to be trusted to perform activities which could affect the health or well-being of their creators. The aim of the following research is to present the autonomous two-level agent framework (ATAF), a framework for intelligent agents to operate within a robotic entity. The entity must be able to adapt to various environments and situations and react in a manner consistent with its environment.

M.Sc. (Computer Science) / This research study investigates multi-agent systems (MASs), artificial life concepts and machine learning, amongst other things, in answering the key research question: “How can a generic multi-agent system integrate with machine learning through artificial life principles?” In answering this question, this dissertation illustrates the design and development of a generic multi-agent, life simulation and learning software framework. This framework simplifies and enables the realisation of MASs in solving complex problems in multiple domains. Finally, this research presents a prototype solution as a proof of concept of the framework’s strengths and weaknesses. The research study illustrates the design of MASs utilising sound design principles, patterns and methodologies. Furthermore, this research explores the requirements for creating and integrating MASs with other technologies, as well as the possible pitfalls in creating such large-scale systems. In addressing the necessity of learning, several machine learning techniques are examined and reinforcement learning is identified as an ideal candidate for the proposed framework. In addition, by understanding the overall machine learning process, the proposed framework integrates machine learning as three separate processes: data extraction, learning and inference. Lastly, the literature study focuses on artificial life, specifically its use in MASs, and defines what constitutes an intelligent system. This research depicts artificial life as a plausible natural integrator between MAS and machine learning technologies. The proposed framework presented in this dissertation consists of five core agent modules that can be extended, depending on the problem domain requirements. The framework in itself is self-containing and independent of any concrete implementation. A multi-agent antivirus system is presented as the prototype implementation of the proposed framework. A quantitative and qualitative analysis was conducted, identifying the results of the prototype and generic framework while highlighting strengths and weaknesses. The contribution of this research is found partly in the proposed generic framework as a means of augmenting mechanisms for MAS design and development by means of artificial life and machine learning integration. In a broader context, this research serves as a foundation towards creating advanced MAS frameworks, leading to numerous interesting and influential agent-oriented applications.

M.Sc. (Computer Science) / This dissertation presents a method for learning from existing building designs and generating new building layouts. Generating fully furnished building layouts could be very useful for video games or for assisting architects when designing new buildings. The core concern is to drastically reduce the workload required to design building layouts. The implemented prototype features a Computer Aided Design system, named CABuilD that allows users to design fully furnished multi-storey building layouts. Building layouts designed using CABuilD can be taught to an Artificial Immune System. The Artificial Immune System tracks information such as building layouts, room sizes and furniture layouts. Once building layouts has been taught to the artificial immune system, a generation algorithm can utilise the information in order to generate fully furnished building layouts. The generation algorithm that is presented allows fully furnished buildings to be generated from high-level information such as the number of rooms to include and a building perimeter. The presented algorithm differs from existing building generation methods in the following ways: Firstly existing methods either ignore building perimeters or assume a buildings perimeter is a rectangle. The presented method allows the user to specify a closed polygon as a building perimeter which will guide the generation of the building layout. Secondly existing generation methods tend to run from a set of rules. The implemented system learns from existing building layouts, effectively allowing it to generate different building types based on the building layouts that were taught to the system. Thirdly, the system generates both the building layout as well as the furniture within rooms. Existing systems only generate the building layout or the furniture, but not both. The prototype that was implemented as a proof of concept uses a number of biologically inspired techniques such as Ant algorithms, Particle Swarm Optimisation and Artificial Immune Systems. The system also employs multiple intelligent agents in order to furnished rooms. The prototype is capable of generating furnished building layouts in merely a few seconds, much faster than a human could design such a layout. Possible improvements and future work is presented at the end of the dissertation.