Search

The theory and practice of evolutionary robotics is well established (Nolfi and Floreano, 2000). However, the overwhelming majority of work in evolutionary robotics has, to date, been concerned with evolving a robot’s control system. The robot’s control system is normally expressed entirely in software: it may be a program, or weights in an artificial neural network (ANN). Then the instantiation of a robot from its genome typically requires only that the control program coded by the genome is uploaded into the fixed robot hardware, ready for fitness evaluation. This is not to belittle the difficult challenges of evolutionary robotics, but to point out that conventional practice has, with a very small number of notable exceptions, avoided the altogether greater challenge of evolving a robot’s physical body.

It has long been assumed that swarm systems are robust, in the sense that the failure of individual robots will have little detrimental effect on a swarm’s overall collective behaviour. However, Bjerknes and Winfield [1] have recently shown that this is not always the case, particularly in the event of partial failures (such as motor failure). The reliability modelling in [1] shows that overall system reliability rapidly decreases with swarm size, therefore this is a problem that cannot simply be solved by adding more robots to the swarm. Instead, future large-scale swarm systems will need an active approach to dealing with failed individuals if they are to achieve a high level of fault tolerance.

Underwater exploration is important for mapping out the oceans, environmental monitoring, and search and rescue, yet water represents one of the most challenging of operational environments. The CoCoRo project proposes to address these challenges using cognitive swarm intelligent systems. We present here CoCoRoSim, an underwater swarm robotics simulation used in designing underwater swarm robotic systems. Collective coordination of robots represents principle challenge here, and use simulation in evaluating shoaling algorithm performance given the communication, localization and orientation challenges of underwater environments. We find communication to be essential for well-coordinated shoals, and provided communication is possible, inexact localization does not significantly impact performance. As a proof of concept simulation is employed in evaluating shoaling performance in turbulent waters.

In this paper, the reliability of a collective robotic system is analysed using two different techniques from the field of reliability engineering. The techniques, Failure Mode and Effect Analysis (FMEA) and Fault Tree Analysis (FTA), are used to analyse and compare two variants of a previously developed ‘autonomous morphogenesis’ controller. The reliability of the controller is discussed and areas where improvements could be made are suggested. The usefulness of FMEA and FTA as aids to the design of fault tolerant collective robotic systems, and the comparative effectiveness of the two approaches, is also discussed.

Link failure and unreachable nodes due to interference from external devices are common problems in WSNs. These interferences can be a major inhibitor to node performance and network stability. In order to tolerate these failures, we propose an immune-inspired self healing system where an individual node can detect degradations in network performance, perform diagnostic tests, and provide automated immediate response to recover the network to a stable state. We evaluate and compare the performance of our approach with other routing protocols on a testbed environment using TelosB hardware motes.

This paper explores approaches to Adaptive Information Filtering (AIF) in the context of changing user interests. Based on the existing artificial immune system for email classification (AISEC), we demonstrate an effective extension to classification based on the body of emails. Widening this to the problem of AIF on dynamic web content, we propose to explore dynamic clonal selection algorithms (DCSAs) that include dynamically changing thresholds.

The PPAM is a hardware architecture for a robust, bidirectional and scalable hetero-associative memory. It is fundamentally different from the traditional processing methods which use arithmetic operations and consequently ALUs. In this paper, we present the results of applying the PPAM to a real-world robotics hand-eye coordination task. A comparison is performed with a nearest neighbour technique that was originally used to associate the same dataset. The number of memory load/store operations and the number of ALU operations for the nearest neighbour algorithm is compared with the corresponding PPAM which acheives the same association. It was determined that 29 conflict resolving nodes were required to fully store and recall the entire dataset and the maximum number of memory locations required in any node was 160, with the average and quartiles being much lower.

We present the design of a new structural extension for the e-puck mobile robot. The extension may be used to transform what is a swarm robotics platform into a self-reconfigurable modular robotic system. As a proof of concept, we present an algorithm for controlling the collective locomotion of a group of e-pucks that are equipped with the extension. Our approach proves itself to be an effective method of coordinating the movement of a group of physically connected e-pucks. Furthermore, the system shows robustness in its ability to self-reconfigure and self-assemble following a disruption which alters the group's structure.

Granuloma formation is a complex process involving a variety of mechanisms acting in concert to afford an inflammatory lesion that is able to contain and destroy intracellular pathogens. While it is crucial for host defence, inappropriate granulomatous inflammation can also damage the host. Granuloma formation is comprised of four main steps : (1) the triggering of T cells by antigen presenting cells, represented by alveolar macrophages and dendritic cells; (2) the release of cytokines and chemokines by macrophages, activated lymphocytes and dendritic cells. Cytokines and chemokines attract and retain in the lung the immuno-inflammatory cell populations in the lung, inducing their survival and proliferation at the site of ongoing inflammation, favouring (3) the stable and dynamic accumulation of immunocompetent cells and the formation of the organised structure of the granuloma. In granulomatous diseases, the last phase (4) of granuloma formation generally ends in fibrosis. Granuloma formation is initiated when an infectious diseases enters the host. Macrophages will ‘eat′ or engulf bacteria to prevent it from spreading. However, bacteria will infect macrophages and duplicate. Despite the fact that macrophages are able to stop the infection, bacteria will use macrophages as a ‘taxi′ to spread the disease within the host leading to cell lysis or the breaking down of the structure of the cell. Infected macrophages then will emit a signal which indicates that they have been infected and this signal will lead other macrophages to move to the site of infection, to form a ring around the infected macrophages thus isolating the infected cells from the uninfected cells. This will finally lead to the formation of a granuloma that represents a chronic inflammatory response initiated by various infectious and non-infectious agents.

In this paper a system which optimises parameter values for the Receptor Density Algorithm (RDA), an algorithm inspired by T-cell signalling, is described. The parameter values are optimised using a genetic algorithm. This system is used to optimise the RDA parameters to obtain the best results when finding anomalies within a large prerecorded dataset, in terms of maximising detection of anomalies and minimising false-positive detections. A trade-off front between the objectives is extracted using NSGA-II as a base for the algorithm. To improve the run-time of the optimisation algorithm with the goal of achieving real-time performance, the system exploits the inherent parallelism of GPGPU programming techniques, making use of the CUDA language and tools developed by NVidia to allow multiple evaluations of a given data set in parallel.

In the past decades, a number of genetic programming techniques have been developed to evolve machine instructions. However, these approaches typically suffer from a lack of scalability that seriously impairs their applicability to real-world scenarios. In this paper, a novel self-scaling instruction generation method is introduced, which tries to overcome the scalability issue by using Cartesian Genetic Programming. In the proposed method, a dual-layer network architecture is created: one layer is used to evolve a series of instructions while the other is dedicated to the generation of loop control parameters.

In this paper, we propose an algorithm for cluster analysis inspired by the lymphocyte-cytokine network in the immune system. Our algorithm attempts to optimally represent a large data set by its principle subset whilst maximising the data kernel density distribution. Experiments show that the output data set created by our algorithm effectively represents the original input data set, according to the Kullback-Leibler divergence metric. We compare the performance of our approach with the well-known aiNet algorithm and find our approach provides a significant improvement on the representation of the final data set.

Experimental autoimmune encephalomyelitis (EAE) is an animal model of human autoimmune diseases in general, and multiple sclerosis (MS) in particular [2]. The animal disease is mediated through a network of cells; encephalitogenic CDThelper (CD4Th1) cells are activated in the peripheral lymph nodes following immunization for EAE, and migrate to the central nervous system where they induce activation of microglia, macrophages and dendritic cells (DCs). The resultant inflammation causes demyelination of neurons, prompting the presentation of myelin basic protein (MBP) to additional encephalitogenic T cell populations in the cervical lymph nodes by migratory DCs. The spontaneous recovery that occurs following autoimmune episodes is associated with the induction of apoptosis in encephalitogenic CD4Th1 cells by a CD8 regulatory T cell (CD8Treg) population [1,6]. Tang et al [7] demonstrated a mechanism where CD4Treg cells assist DCs in priming CD8Treg cells, which mediate down-regulation of the autoimmune response through selective apoptotic elimination of CD4Th1 cells. The exact location of this down-regulation of the autoimmune response by CD8Treg cells is unknown. One possibility is that DCs migrating from the central nervous system, potentially carrying both MBP (from neurons) and T cell receptor (TCR) derived peptides (Fr3 and CDR1/2; from encephalitogenic CD4Th1 cells), are simultaneously priming both CD4Th1 and Treg cell populations, and that regulation of autoimmunity is taking place around the DC. In this paper we investigate the effect of regulation of CD4Th1 cells by CD8Treg cells through use of a simulation (ARTIMUS) we have developed to support investigation into EAE [5]. We investigate the hypothesis that a significant portion of CD4Th1 regulation by CD8Treg cells occurs around DCs that prime both T cell populations, and hence that regulation occurring around such DCs is critical for recovery from disease.

Specialised tissues of the immune system including lymph nodes, tonsils, spleen and Peyer’s Patches have key roles in the initiation of adaptive immune responses to infection. To understand the molecular mechanisms involved in the development of this tissue, mice deficient for key genes in this process have been developed and analysed, leading to a basic model describing tissue formation. Although this approach has provided some key insights into the molecular mechanisms involved, due to the complexity of gene expression patterns it has not been possible to fully understand the process of lymphoid tissue organogenesis.

Leishmania donovani is an obligate intracellular parasite responsible for the systemic disease visceral leishmaniasis. During the course of the disease, the parasite is found in the spleen, liver and bone marrow. Characteristic of the liver immune response to leishmaniasis is a type of inflammation (“ggranulomatous inflammation”) that results in the formation of granulomas, structures comprised of an infiltrate of mononuclear cells surrounding a core of infected macrophages. Granulomas help limit the spread of infection and facilitate the killing of parasites.

The biological term Homeostasis refers to an organisms ability to maintain steady states of operation in a massively changing internal and external environment. Pervasive Adaptive (PerAda) systems operate in potentially hostile environments where there is a need to cope with a wide variety of changes, over multiple timescales, deal with interacting agents, cope with degraded data and cope with failing components. In the context of this paper, we consider the case of a swarm, and organism, of robotic units, as those developed as part of the SYMBRION project. Such a collective robotic system can be considered a case of a PerAda system.

This paper presents the novel use of the Neural-endocrine architecture for swarm robotic systems. We make use of a number of behaviours to give rise to emergent swarm behaviour to allow a swarm of robots to collaborate in the task of foraging. Results show that the architecture is amenable to such a task, with the swarm being able to successfully complete the required task.

This paper presents a high level dynamic task allocation algorithm that is inspired by the biological development process and the immune system. For a microprocessor (μP) array, a program is partitioned into a number of workload oriented tasks with data dependencies and a number of internal status-oriented tasks. Each μP in the array is capable of processing one of these tasks. The algorithm assigns tasks to the μP array that satisfies the requirements of the problem, and it dynamically recovers the system from faults at runtime.

In this paper, we present the first stage of our research strategy to develop an immune-inspired solution for detecting anomalies in a foraging swarm robotic system with an immuno-engineering approach. Within immuno-engineering, the initial stage of our research involves the understanding of problem domain, namely anomaly detection, in a foraging swarm robotic system deployed in dynamic environments. We present a systematically derived set of activities for this stage derived with Goal Structuring Notation and results of experiments carried out to establish the time-varying behaviour and how anomalies manifest themselves. Our future work will then be used to select and tailor an appropriate AIS algorithm to provide an effective and efficient means of anomaly detection.

Experimental Autoimmune Encephalomyelitis (EAE) is an autoimmune disease in mice which serves as a model for multiple sclerosis in humans [4,5]. The disease constitutes the direction of immunity towards myelin, an insulatory material that covers neurons. The consequential damage to the central nervous system (CNS) can lead to paralysis and death [6].

The T cell is able to perform fine-grained anomaly detection via its T Cell Receptor and intracellular signalling networks. We abstract from models of T Cell signalling to develop a new Artificial Immune System concepts involving the internal components of the TCR. We show that the concepts of receptor signalling have a natural interpretation as Parzen Window Kernel Density Estimation applied to anomaly detection. We then demonstrate how the dynamic nature of the receptors allows anomaly detection when probability distributions vary in time.

This paper proposes the hardware implementation of a novel evolutionary algorithm inspired by protein/substrate binding exploited in artificial immune networks. The immune network inspired evolutionary algorithm has been developed in direct response to an application in clinical neurology, the diagnosis of Parkinson's Disease, but is now being considered for other more demanding applications where real-time processing of data is required. The inspiration for the algorithm and its proposed implementation in hardware is presented. The effectiveness of the approach is shown by results obtained from a software implementation.

This paper addresses the classification task of data mining (a form of supervised learning) in the context of an important bioinformatics problem, namely the prediction of protein functions. This problem is cast as a hierarchical classification problem, where the protein functions to be predicted correspond to classes that are arranged in a hierarchical structure, in the form of a class tree. The main contribution of this paper is to propose a new Artificial Immune System that creates a new representation for proteins, in order to maximize the predictive accuracy of a hierarchical classification algorithm applied to the corresponding protein function prediction problem.

The Tunable Activation Threshold hypothesis of T Cells is investigated through computational modelling of T cell signalling pathways. Modelling techniques involving the i¾?-calculus and the PRISM model checker are presented, and are applied to produce a stochastic model of T cell signalling. Initial results which demonstrate tuning of T cells are presented.

The adaptable lymphocyte hypothesis is identified as a possible source of inspiration for artificial immune systems. Based on a number of qualitative investigations we identify some properties of a theoretical system (the tunable activation threshold model and excitability) that could be applicable in an engineering domain. An example is shown of how we could exploit these properties.

We present an empirical investigation of a mathematical model of cytokine and cell interactions, as derived by Hone and van den Berg, with the intention of ascertaining its potential for deployment in an engineering context. The model's behavioural robustness regarding various patterns of input is assessed, as is it's scope for manipulation through parameter adjustments. Based on these observations we make an addition to the system and attempt to deliberately engineer particular behavioural properties. Finally, the system's suitability for use in engineering is assessed.

Many electronic systems would benefit from the inclusion of self-regulatory mechanisms. We strive to build systems that can predict, or be aware of, imminent threats upon their specified operation. Then, based on this prediction, the system can alter its operation or configuration to circumvent the effects of the threat. In this position paper, we discuss the role of the immune system can play in serving as inspiration for the development of homeostatic engineered systems, through the development of an immune inspired extensible architecture. We outline the major requirements for such an architecture, and discuss issues that arise as a result and propose possible solutions: things are never as simple as they first appear.

We present an Immune Inspired Algorithm, based on CLONALG, for software test data evolution. Generated tests are evaluated using the mutation testing adequacy criteria, and used to direct the search for new tests. The effectiveness of this algorithm is compared against an elitist Genetic Algorithm, with effectiveness measured by the number of mutant executions needed to achieve a specific mutation score. Results indicate that the Immune Inspired Approach is consistently more effective than the Genetic Algorithm, generating higher mutation scoring test sets in less computational expense.

In this position paper, we argue that the field of Artificial Immune Systems (AIS) has reached an impass. For many years, immune inspired algorithms, whilst having some degree of success, have been limited by the lack of theorectical advances, the adoption of a limited immune inspired approach and the limited application of AIS to hard problems.

This paper highlights degeneracy as being an important property in both the immune system and biology in general. From this, degeneracy is chosen as a candidate to inspire artificial immune systems. As a first step in exploiting the power of degeneracy, we follow the conceptual framework approach and build an abstract computational model in order to understand the properties of degenerate detectors free of any application bias. The model we build is based on the activation of TH cell in the lymph node, as lymph nodes are the sites in the body where the adaptive immune response to foreign antigen in the lymph are activated. The model contains APC, antigen and TH cell agents that move and interact in a 2-dimensional cellular space. The TH cell agent receptors are assumed to be degenerate and their response to different antigen agents is measured. Initial observations and results of our model are presented and highlight some of the possibilities of degenerate detector recognition.

When constructing biologically inspired algorithms, important properties to consider are openness, diversity, interaction, structure and scale. In this paper, we focus on the property of diversity. Introducing diversity into biologically inspired paradigms is a key feature of their success. Within the field of Artificial Immune Systems, little attention has been paid to this issue. Typically, techniques of diversity introduction, such as simple random number generation, are employed with little or no consideration to the application area. Using function optimisation as a case study, we propose a simple immune inspired mutation operator that is tailored to the problem at hand. We incorporate this diversity operator into a well known immune inspired algorithm, aiNet. Through this approach, we show that it is possible to improve the search capability of aiNet on hard to locate optima. We further illustrate that by incorporating the same mutation operator into aiNet when applied to clustering, it is observed that performance is neither improved nor sacrificed.

An exact Markov chain model of the B-cell algorithm (BCA) is constructed via a novel possible transit method. The model is used to formulate a proof that the BCA is convergent absolute under a very broad set of conditions. Results from a simple numerical example are presented, we use this to demonstrate how the model can be applied to increase understanding of the performance of the BCA in optimizing function landscapes as well as giving insight into the optimal parameter settings for the BCA.

In this conceptual paper, we consider the state of artificial immune system (AIS) design today, and the nature of the immune theories on which they are based. We highlight the disagreement amongst many immunologists regarding the concept of self–non-self discriminations in the immune system, and go on describe on such model that removes altogether the requirement for self–non-self discrimination. We then identify the possible inspiration ideas for AIS that can be gained from such new, and often radical, models of the immune system. Next, we outline a possible approach to designing AIS that are inspired by new immune theories, following a suitable methodology and selecting appropriate modelling tools. Lastly, we follow our approach and present an example of how the AIS designer might take inspiration from a specific property of a new immune theory. This example highlights our proposed method for inspiring the design of the next generation of AIS.

Conceptual Track.- Effect of Synthetic Emotions on Agents' Learning Speed and Their Survivability.- From the Inside Looking Out: Self Extinguishing Perceptual Cues and the Constructed Worlds of Animats.- Globular Universe and Autopoietic Automata: A Framework for Artificial Life.- May Embodiment Cause Hyper-Computation?.- Perception as a Dynamical Sensori-Motor Attraction Basin.- Toward Genuine Continuity of Life and Mind.- Morphogenesis and Development.- Biological Development of Cell Patterns: Characterizing the Space of Cell Chemistry Genetic Regulatory Networks.- A Coarse-Coding Framework for a Gene-Regulatory-Based Artificial Neural Tissue.- A Computational Model of Cellular Morphogenesis in Plants.- A Developmental Model for Generative Media.- Evolutionary Simulations of Maternal Effects in Artificial Developmental Systems.- METAMorph: Experimenting with Genetic Regulatory Networks for Artificial Development.- Morphological Plasticity: Environmentally Driven Morphogenesis.- A Self-organising, Self-adaptable Cellular System.- Self-repair Ability of a Toroidal and Non-toroidal Cellular Developmental Model.- Simulating Evolution with a Computational Model of Embryogeny: Obtaining Robustness from Evolved Individuals.- Topology Changes Enable Reaction-Diffusion to Generate Forms.- Robotics and Autonomous Agents.- Aggregation Behaviour as a Source of Collective Decision in a Group of Cockroach-Like-Robots.- (Co)Evolution of (De)Centralized Neural Control for a Gravitationally Driven Machine.- Co-evolution of Structures and Controllers for Neubot Underwater Modular Robots.- CoEvolutionary Incremental Modelling of Robotic Cognitive Mechanisms.- A Dynamical Systems Approach to Learning: A Frequency-Adaptive Hopper Robot.- An Evolved Agent Performing Efficient Path Integration Based Homing and Search.- Evolving Neural Mechanisms for an Iterated Discrimination Task: A Robot Based Model.- Hysteresis and the Limits of Homeostasis: From Daisyworld to Phototaxis.- Is an Embodied System Ever Purely Reactive?.- t for Two Linear Synergy Advances the Evolution of Directional Pointing Behaviour.- Self-assembly on Demand in a Group of Physical Autonomous Mobile Robots Navigating Rough Terrain.- Superlinear Physical Performances in a SWARM-BOT.- Timescale and Stability in Adaptive Behaviour.- Whisker-Based Texture Discrimination on a Mobile Robot.- Evolutionary Computation and Theory.- Analysing the Evolvability of Neural Network Agents Through Structural Mutations.- Coevolutionary Species Adaptation Genetic Algorithms: A Continuing SAGA on Coupled Fitness Landscapes.- Evolution and the Regulation of Environmental Variables.- Evolutionary Transitions as a Metaphor for Evolutionary Optimisation.- Genetic Assimilation and Canalisation in the Baldwin Effect.- How Do Evolved Digital Logic Circuits Generalise Successfully?.- How Niche Construction Can Guide Coevolution.- Measuring Diversity in Populations Employing Cultural Learning in Dynamic Environments.- On a Quantitative Measure for Modularity Based on Information Theory.- On the Mean Convergence Time of Multi-parent Genetic Algorithms Without Selection.- The Quantitative Law of Effect is a Robust Emergent Property of an Evolutionary Algorithm for Reinforcement Learning.- Self-adaptation of Genome Size in Artificial Organisms.- Cellular Automata.- An Architecture for Modelling Emergence in CA-Like Systems.- The Density Classification Problem for Multi-states Cellular Automata.- Evolving Cellular Automata by 1/f Noise.- Evolving Sequential Combinations of Elementary Cellular Automata Rules.- Penrose Life: Ash and Oscillators.- Playing a 3D Game of Life in an Interactive Virtual Sandbox.- Using Dynamic Behavior Prediction to Guide an Evolutionary Search for Designing Two-Dimensional Cellular Automata.- Models of Biological Systems and Their Applications.- CelloS: A Multi-level Approach to Evolutionary Dynamics.- A Cytokine Formal Immune Network.- Examining Refuge Location Mechanisms in Intertidal Snails Using Artificial Life Simulation Techniques.- A Method for Designing Ant Colony Models and Two Examples of Its Application.- Simulating Artificial Organisms with Qualitative Physiology.- Slime Mould and the Transition to Multicellularity: The Role of the Macrocyst Stage.- Ant Colony and Swarm Systems.- Ant Clustering Embeded in Cellular Automata.- Ant-Based Computing.- Collective Behavior Analysis of a Class of Social Foraging Swarms.- Evolving Annular Sorting in Ant-Like Agents.- Flocking Control of Multiple Interactive Dynamical Agents with Switching Topology via Local Feedback.- Evolution of Communication.- Cultural and Biological Evolution of Phonemic Speech.- Grammar Structure and the Dynamics of Language Evolution.- Interactions Between Learning and Evolution in Simulated Avian Communication.- Perceptually Grounded Lexicon Formation Using Inconsistent Knowledge.- Simulation of Social Interactions.- Artificial Life Meets Anthropology: A Case of Aggression in Primitive Societies.- Emergence of Structure and Stability in the Prisoner's Dilemma on Networks.- Multi-agent-based Simulation for Formation of Institutions on Socially Constructed Facts.- Self-replication.- Extensions and Variations on Construction of Autoreplicators in Typogenetics.- The Good Symbiont.- Self-description for Construction and Execution in Graph Rewriting Automata.- Artificial Chemistry.- Artificial Metabolic System: An Evolutionary Model for Community Organization in Metabolic Networks.- Explicit Collision Simulation of Chemical Reactions in a Graph Based Artificial Chemistry.- Self-replication and Evolution of DNA Crystals.- Posters.- All Else Being Equal Be Empowered.- Artificial Homeostatic System: A Novel Approach.- Artificial Life for Natural Language Processing.- A Co-evolutionary Epidemiological Model for Artificial Life and Death.- CoEvolution of Effective Observers and Observed Multi-agents System.- Comparative Reproduction Schemes for Evolving Gathering Collectives.- Construction-Based and Inspection-Based Universal Self-replication.- Coordinating Dual-Mode Biomimetic Robotic Fish in Box-Pushing Task.- Effects of Spatial Growth on Gene Expression Dynamics and on Regulatory Network Reconstruction.- Evolution of Song Communication in a 2D Space.- A Fitness-Landscape for the Evolution of Uptake Signal Sequences on Bacterial DNA.- The Genetic Coding Style of Digital Organisms.- Growing Biochemical Networks: Identifying the Intrinsic Properties.- Multi-population Cooperative Particle Swarm Optimization.- On Convergence of Dynamic Cluster Formation in Multi-agent Networks.- On the Unit of Selection in Sexual Populations.- Periodic Motion Control by Modulating CPG Parameters Based on Time-Series Recognition.- Self-organized Criticality on Growing Scale-Free Networks.- Synapsing Variable Length Crossover: An Algorithm for Crossing and Comparing Variable Length Genomes.- Valency for Adaptive Homeostatic Agents: Relating Evolution and Learning.

Many researchers are developing frameworks inspired by natural, especially biological, systems to solve complex real-world problems. This work extends previous work in the field of biologically inspired computing, proposing an artificial endocrine system for autonomous robot navigation. Having intrinsic self-organizing behaviour, the novel artificial endocrine system can be applied to a wide range of problems, particularly those that involve decision making under changing environmental conditions, such as autonomous robot navigation. This work draws on “embodied cognitive science”, including the study of intelligence, adaptivity, homeostasis, and the dynamic aspects of cognition, in order to help lay down fundamental principles and techniques for a novel approach to more biologically plausible artificial homeostatic systems. Results from using the artificial endocrine system to control a simulated robot are presented.

The immune system is highly distributed, highly adaptive, self-organising in nature, maintains a memory of past encounters and has the ability to continually learn about new encounters. From a computational point of view, the immune system has much to offer by way of inspiration to computer scientists and engineers alike. As computational problems become more complex, increasingly, people are seeking out novel approaches to these problems, often turning to nature for inspiration. A great deal of attention is now being paid to the vertebrae immune system as a potential source of inspiration, where it is thought that different insights and alternative solutions can be gleaned, over and above other biologically inspired methods. Given this rise in attention to the immune system, it seems appropriate to explore this area in some detail. This survey explores the salient features of the immune system that are inspiring computer scientists and engineers to build Artificial Immune Systems (AIS). An extensive survey of applications is presented, ranging from network security to optimisation and machine learning. However, this is not complete, as no survey ever is, but it is hoped this will go some way to illustrate the potential of this exciting and novel area of research.

Proteins are the driving force in development (embryogenesis) and the immune system. Here we describe how a model of proteins designed for evolutionary development in computers can be combined with a model of immune systems. Full details of a prototype system are provided, and preliminary experiments presented. Results show that evolution is able to adjust the mapping between input data and antigens and cause useful changes to the subnetworks formed by the immune algorithm.

Artificial Immune Systems (AIS) have recently been proposed as an additional soft computing paradigm. This paper proposes a new multi-layered unsupervised machine learning algorithm inspired by the vertebrate immune system. The algorithm has been tested on benchmark data and has shown a great deal of potential for data reduction and clustering tasks. This paper presents an overview of the algorithm, drawing analogies to the vertebrae immune system where appropriate. Results are presented for three data sets and observations are made about the potential for adapting the algorithm for a continuous learning paradigm.

Verifying the published results of algorithms is part of the usual research process. This helps to both validate the existing literature, but also quite often allows for new insights and augmentations of current systems in a methodological manner. This is very pertinent in emerging new areas such as Artificial Immune Systems, where it is essential that any algorithm is well understood and investigated. The work presented in this paper results from an investigation into the opt-aiNET algorithm, a well-known immune inspired algorithm for function optimisation. Using the original source code developed for opt-aiNET, this paper identifies two minor errors within the code, propose a slight augmentation of the algorithm to automate the process of peak identification: all of which affect the performance of the algorithm. Results are presented for testing of the existing algorithm and in addition, for a slightly modified version, which takes into account some of the issues discovered during the investigations.

Over time programming languages develop, paradigmsevolve, development teams change. The effect of this is that test suites wear out, therefore these also need to evolve. Mutation testing is an effective fault-based testing approach, but it is computationally expensive. Any evolutionary based approach to this process needs to simultaneously manage execution costs. In this conceptual paper we adopt immune systems as a metaphor for the basis of an alternative mutation testing system. It is envisaged that through monitoring of the development environment, a minimal set of effective mutations and test cases can be developed – a ‘vaccine’ – that can be applied to the software development process to protect it from errors – from infections.

Since their development, AIS have been used for a number of machine learning tasks including that of classification. Within the literature, there appears to be a lack of appreciation for the possible bias in the selection of various representations and affinity measures that may be introduced when employing AIS in classification tasks. Problems are then compounded when inductive bias of algorithms are not taken into account when applying seemingly generic AIS algorithms to specific application domains. This paper is an attempt at highlighting some of these issues. Using the example of classification, this paper explains the potential pitfalls in representation selection and the use of various affinity measures. Additionally, attention is given to the use of negative selection in classification and it is argued that this may be not an appropriate algorithm for such a task. This paper then presents ideas on avoiding unnecessary mistakes in the choice and design of AIS algorithms and ultimately delivered solutions.

Within immunology, new theories are constantly being proposed that challenge current ways of thinking. These include new theories regarding how the immune system responds to pathogenic material. This conceptual paper takes one relatively new such theory: the Danger theory, and explores the relevance of this theory to the application domain of web mining. Central to the idea of Danger theory is that of a context dependant response to invading pathogens. This paper argues that this context dependency could be utilised as powerful metaphor for applications in web mining. An illustrative example adaptive mailbox filter is presented that exploits properties of the immune system, including the Danger theory. This is essentially a dynamical classification task: a task that this paper argues is well suited to the field of artificial immune systems, particularly when drawing inspiration from the Danger theory.

When considering function optimisation, there is a trade off between quality of solutions and the number of evaluations it takes to find that solution. Hybrid genetic algorithms have been widely used for function optimisation and have been shown to perform extremely well on these tasks. This paper presents a novel algorithm inspired by the mammalian immune system, combined with a unique mutation mechanism. Results are presented for the optimisation of twelve functions, ranging in dimensionality from one to twenty. Results show that the immune inspired algorithm performs significantly fewer evaluations when compared to a hybrid genetic algorithm, whilst not sacrificing quality of the solution obtained.

Within immunology, new theories are constantly being proposed that challenge current ways of thinking. These include new theories regarding how the immune system responds to pathogenic material. This conceptual paper takes one relatively new such theory: the Danger theory, and explores the relevance of this theory to the application domain of web mining. Central to the idea of Danger theory is that of a context dependant response to invading pathogens. This paper argues that this context dependency could be utilised as powerful metaphor for applications in web mining. An illustrative example adaptive mailbox filter is presented that exploits properties of the immune system, including the Danger theory. This is essentially a dynamical classification task: a task that this paper argues is well suited to the field of artificial immune systems, particularly when drawing inspiration from the Danger theory.

This chapter Enscribes a machine learning system based on metaphors taken from the human immune system. This learning system, known as an Artificial Immune System (AIS), has been Enveloped over the past 3 years. The current implementation,Jisys,embodies the results of this research. However, the Jisys implementation requires further Envelopment as well as application to complex real world problems. This chapter Enscribes future Envelopments ofJisysas well as consiEnration of how it can be applied to a complex problem in the domain of mortgage fraud Entection. It should not be read as a Ensign document, although it contains elements of such a document, rather it should be read as an indication of the directions which need to be followed, the issues which need to be addressed and some suggested solutions

This paper discusses the notion of “core bio-inspired services” - low-level services providing basic bio-inspired mechanisms, such as evaporation, aggregation or spreading - shared by higher-level services or applications. Design patterns descriptions of self-organising mechanisms, such as gossip, morphogenesis, or foraging, show that these higher-level mechanisms are composed of basic bio-inspired mechanisms (e.g. digital pheromone is composed of spreading, aggregation and evaporation). In order to ease design and implementation of self-organising applications (or high-level services), by supporting reuse of code and algorithms, this paper proposes BIO-CORE, an execution model that provides these low-level services at the heart of any middleware or infrastructure supporting such applications, and provides them as “core” built-in services around which all other services are built.