Self-organizing Systems exhibit numerous advantages such as robustness, adaptivity and scalability, and thus provide a solution for the increasing complexity we face within technical systems. While they are attractive solutions, due to their nature, designing self-organizing systems is not a straightforward task. Artificial evolution has been proposed as a possible way to build self-organizing systems, but there are still many open questions on how an engineer should apply this method for this purpose. In this paper we propose a design methodology for evolving self-organizing systems, that marks the major cornerstones and decisions the designer has to face, thus providing a practical set of guidelines.
Wireless Sensor Networks (WSNs) led the way to new forms of communications, which extend today the Internet paradigm to unforeseen boundaries. The legacy industry, however, is slower to adopt this technology, mainly for security reasons. Self-managed security systems allowing a quicker detection of and better resilience to attacks, may counterbalance this reluctance. We propose in this paper a hybrid threat detection and security adaptation system, designed to run on top of industrial WSNs. We explain why this system is suitable for architectures mainly composed of constrained or sleeping devices, while being able to achieve a fair level of autonomous security.
This paper introduces a distributed, self-organizing approach to load control based on voltage measurement. A local voltage measurement defines a Level of Service (LoS), which is balanced with the neighboring households in order to avoid extreme restrictions of energy use in a single household. The approach distinguishes four criticality classes for devices which suspend themselves at specific LoS thresholds.
This paper seeks to systematically explore the efficiency of the uncoordinated information-assisted parking search in urban environments with two types of parking resource facilities: inexpensive but limited facilities (public) and expensive yet unlimited ones (private); an additional cruising cost is incurred when deciding for a public facility but failing to actually utilize one. Drivers decide whether to go for the public or directly for the private facilities, assuming perfect knowledge of prices and costs, total parking capacities and demand; the latter information can be broadcast by an ideal centralized information dissemination mechanism, assisting the otherwise uncoordinated parking search process. Drivers are viewed as strategic decision-makers that aim at minimizing the cost of the acquired parking spot. We formulate the resulting game as an instance of resource selection games and derive its Nash equilibria and their dependence on the environmental parameters such as the parking demand and supply as well as the pricing policy. The cost at the equilibrium states is compared to that under the optimal resource assignment (dictated to the drivers directly by an ideal centralized scheme) and conditions are derived for minimizing the related price of anarchy. Finally, the numerical results and the presented discussion provide hints for the practical management and pricing of public and private parking resources.
This work analyses the practice of sister city pairing. We investigate structural properties of the resulting city and country networks and present rankings of the most central nodes in these networks. We identify different country clusters and find that the practice of sister city pairing is not influenced by geographical proximity but results in highly assortative networks.
There is an emerging research area to adopt bio-inspired algorithms to self-organize an information network system. Despite strong interests on their benefits, i.e. high robustness, adaptability, and scalability, the behavior of bio-inspired algorithms under non-negligible perturbation such as loss of information and failure of nodes observed in the realistic environment is not well investigated. Because of lack of knowledge, none can clearly identify the range of application of a bio-inspired algorithm to challenging issues of information networks. Therefore, to tackle the problem and accelerate researches in this area, we need to understand characteristics of bio-inspired algorithms from the perspective of network control. In this paper, taking a response threshold model as an example, we discuss the robustness and adaptability of bio-inspired model and its application to network control. Through simulation experiments and mathematical analysis, we show an existence condition of the equilibrium state in the lossy environment. We also clarify the influence of the environmental condition and control parameters on the transient behavior and the recovery time.
The opportunistic data dissemination problem for mobile devices is an open topic that has attracted many investigations so far. At the best of our knowledge, none of these approaches takes into account the semantic side of the data shared in an opportunistic network. In this paper, we present an algorithm that, starting from the semantic data annotations given by the users themselves, builds a semantic network representation of the information. Exploiting this description, we detail how two different semantic networks can interact upon contact, in order to spread and receive useful information. In order to provide a performance evaluation of such a solution, we show a preliminary set of results obtained in a simulated scenario.
Wireless sensor networks are attracting increasing interest but suffer from severe challenges such as power constraints and low data reliability. Sensors are often energy-hungry and cannot operate over the long term, and the data they gather are frequently erroneous in complex ways. The two problems are linked, but existing work typically treats them independently: in this paper we consider both side-by-side, and propose a self-organising solution for model-based data collection that reduces errors and communications in a unified fashion.
Agents in open self-organizing systems have to cope with a variety of uncertainties. In order to increase their utility and to ensure stable operation of the overall system, they have to capture and adapt to these uncertainties at runtime. This can be achieved by formulating an expectancy of the behavior of others and the environment. Trust has been proposed as a concept for this purpose. In this paper, we present trust-based scenarios as an enhancement of current trust models. Trust-based scenarios represent stochastic models that allow agents to take different possible developments of the environment's or other agents' behavior into account. We demonstrate that trust-based scenarios significantly improve the agents' capability to predict future behavior with a distributed power management application.
In recent times, the research field of language dynamics has focused on the investigation of language evolution, dividing the work in three evolutive steps, according to the level of complexity: lexicon, categories and grammar. The Naming Game is a simple model capable of accounting for the emergence of a lexicon, intended as the set of words through which objects are named. We introduce a stochastic modification of the Naming Game model with the aim of characterizing the emergence of a new language as the result of the interaction of agents. We fix the initial phase by splitting the population in two sets speaking either language A or B. Whenever the result of the interaction of two individuals results in an agent able to speak both A and B, we introduce a finite probability that this state turns into a new idiom C, so to mimic a sort of hybridization process. We study the system in the space of parameters defining the interaction, and show that the proposed model displays a rich variety of behaviours, despite the simple mean field topology of interactions.
In many networks with distributed operation and self-organization features, acquiring their global topological information is impractical, if feasible at all. Internet protocols drawing on node centrality indices may instead approximate them with their egocentric counterparts, computed out over the nodes' ego-networks. Surprisingly, however, in router-level topologies the approximative power of localized ego-centered measurements has not been systematically evaluated. More importantly, it is unclear how to practically interpret any positive correlation found between the two centrality metric variants. The paper addresses both issues using different datasets of ISP network topologies. We first assess how well the egocentric metrics approximate the original sociocentric ones, determined under perfect network-wide information. To this end we use two measures: their rank-correlation and the overlap in the top-k node lists the two centrality metrics induce. Overall, the rank-correlation is high, in the order of 0.8-0.9, and, intuitively, becomes higher as we relax the ego-network definition to include the ego's r-hop neighborhood. On the other hand, the top-k node overlap is low, suggesting that the high rank-correlation is mainly due to nodes of lower rank. We then let the node centrality metrics drive elementary network operations, such as local search strategies. Our results suggest that, even under high rank-correlation, the locally-determined metrics can hardly be effective aliases for the global ones. The implication for protocol designers is that rank-correlation is a poor indicator for the approximability of centrality metrics.
In this paper we introduce a novel model that can account for the spread of extreme opinions in a human population as a purely local, self-organising process. Our starting point is the well-known and influential Relative Agreement (RA) model of opinion dynamics introduced by Deffuant et al. (2002). The RA model explores the dynamics of opinions in populations that are initially seeded with some number of "extremist" individuals, who hold opinions at the far ends of a continuous spectrum of opinions represented in the abstract RA model as a real value in the range [-1.0, +1.0]; but where the majority of the individuals in the population are, at the outset, " moderates", holding opinions closer to the central mid-range value of 0.0. Various researchers have demonstrated that the RA model generates opinion dynamics in which the influence of the extremists on the moderates leads, over time, to the distribution of opinion values in the population converging to attractor states that can be qualitatively characterised as one of either uni-polar and bi-polar extremes, or reversion to the centre ("central convergence"). However, a major weakness of the RA model is that it pre-supposes the existence of extremist individuals, and hence says nothing to answer the question of " where do extremists come from?" In this paper, we introduce the Relative Disagreement (RD) model, in which extremist individual arise spontaneously and can then exert influence over moderates, forming large groups of polar extremists, via an entirely internal, self-organisation process. We demonstrate that the RD model can readily exhibit the uni-polar, bi-polar, and central-convergence attractors that characterise the dynamics of the RA model, and hence this is the first paper to describe an opinion dynamic model in which extremist positions can spontaneously arise and spread in a population via a self-organising process where opinion-influencing interactions between any two individuals are characterised not only by the extent to which they agree, but also by the extent to which they disagree.
MANETs (mobile ad hoc networks) operate in a self-organised and decentralised way. Attacks against nodes that are highly relied to relay traffic could result in a wide range of service outage. A comprehensive model that could enhance the understanding of network behaviour under attacks is important to the design and construction of resilient self-organising networks. Previously, we modelled MANETs as an aggregation of time-varying graphs into a static weighted graph, in which the weights represent link availability of pairwise nodes. Centrality metrics were used to measure node significance but might not always be optimal. In this paper, we define a new metric called criticality that can capture node significance more accurately than centrality metrics. We demonstrate that attacks based on criticality have greater impact on network performance than centrality-based attacks in real-time MANETs.
Future electrical power networks will be composed of large collections of autonomous components. Self-organization is an organizational concept that promises robust systems with the ability to adapt themselves to system perturbations and failures and thus may yield highly robust systems with the ability to scale freely to almost any size. In this position paper the authors describe the well-established process of use case based derivation of non-functional requirements in energy systems and propose a mapping strategy for aligning properties of self-organizing systems with the ICT- and automation system requirements. It is the strong belief of the authors that such a mapping will be a key factor in creating acceptance of and establishing self-organization in the domain of electrical energy systems.
Cloud computing is an increasingly popular platform for both industry and consumers. The cloud presents a number of unique security issues, such as a high level of distribution and system homogeneity, which require special consideration. In this paper we introduce a resilience architecture consisting of a collection of self-organising resilience managers distributed within the infrastructure of a cloud. More specifically we illustrate the applicability of our proposed architecture under the scenario of malware detection. We describe our multi-layered solution at the hypervisor level of the cloud nodes and consider how malware detection can be distributed to each node.
We investigate different analysis methods for traffic data. The measure for emergence can be used to identify global dependencies in data sets. The measure for target orientation can be used to identify dangerous situations in traffic. We apply these measures in a use case on a data set of the M30 highway in Madrid. The evaluation shows that the measures can be used to predict or to identify abnormal events like accidents in traffic by an evaluation of velocity data or flow data measured by detectors at the road. Such events leads to a decrease of the measures of emergence and target orientation.
The general aim of this paper is to introduce the notion of phase transitions into wireless networking optimization. Although the theory of phase transitions from statistical physics has been employed in optimization theory, phase transitions in the context of optimization of wireless networks have not yet been considered. In wireless networking optimization, given one or more optimization objectives we often need to define mathematically an optimization task, so that a set of requirements is not violated. However, especially recent trends in wireless communications, such as self-organized networks, femto-cellular systems, and cognitive radios, calls for optimization approaches that can be implemented in a distributed and decentralized fashion. Thus we are interested to find utility-based approaches that can be practically employed in a self-organizing network. We argue that phase transitions can be identified and taken appropriately into account in order to eliminate the emergence of undesirable solutions that lie near the point where the phase transition occurs. As an example we present a simple power control problem for a macrocell-femtocell network scenario. We formulate a distributed framework of the problem where we model a phase transition effect by means of a dummy variable in order to exclude solutions lying in the one side of the phase transition.
In this paper three computational models for the study of the evolution of cooperation under cultural propagation are studied: Kin Selection, Direct Reciprocity and Indirect Reciprocity. Two analyzes are reported, one comparing their behavior between them and a second one identifying the impact that different parameters have in the model dynamics. The results of these analyzes illustrate how game transitions may occur depending of some parameters within the models and also explain how agents adapt to these transitions by individually choosing their attachment to a cooperative attitude. These parameters regulate how cooperation can self-organize under different circumstances. The emergence of the evolution of cooperation as a result of the agent's adapting processes is also discussed.
This paper applies basic economic principles which have been developed in financial markets to a future smart grid scenario. Our method allows for autonomous bidding for electricity units to create an emerging market price for electricity. We start with replicating the popular Zero-Intelligence-Plus algorithm and setting it in a electricity supplier-consumer scenario. We identify significant weaknesses of applying this in an electricity market especially when intermittent sources of energy are present or when the supplier to consumer ratio is very small. A new algorithm (ZIP-260) is proposed which includes a measure of fairness based on minimising the deviation across all un-matched demand for a given period. This approach means that no consumer in the system is constantly experiencing an electricity supply deficit. We show and explain how market conditions can lead to collective bargaining of consumers and monopolistic behaviour of suppliers and conclude with observations on automated trading for smart grids.
Future energy grids will need to cope with a multitude of new, dynamic situations. Having sufficient information about energy usage patterns is of paramount importance for the grid to react to changing situations and to make the grid 'smart'. We present preliminary results from an investigation on whether autonomic adaptation of intervals with which individual smart meters report their meter readings can be more effective than commonly used static configurations. A small reporting interval provides close to real-time knowledge about load changes and thus gives the opportunity to balance the energy demand amongst consumers rather than 'burning' surplus capacities. On the other hand, a small interval results in a waste of processing power and bandwidth in case of customers that have rather static energy usage behaviour. Hence, an ideal interval cannot be predicted a priori, but needs to be adapted dynamically. We provide an analytical investigation of the effects of autonomic management of smart meter reading intervals, and we make some recommendations on how this scheme can be implemented.
