The data privacy constraint in online continual learning (OCL), where the data can be seen only once, complicates the catastrophic forgetting problem in streaming data. A common approach applied by the current SOTAs in OCL is with the use of memory saving exemplars or features from previous classes to be replayed in the current task. On the other hand, the prompt-based approach performs excellently in continual learning but with the cost of a growing number of trainable parameters. The first approach may not be applicable in practice due to data openness policy, while the second approach has the issue of throughput associated with the streaming data. In this study, we propose a novel prompt-based method for online continual learning that includes 4 main components: (1) single light-weight prompt generator as a general knowledge, (2) trainable scaler-and-shifter as specific knowledge, (3) pre-trained model (PTM) generalization preserving, and (4) hard-soft updates mechanism. Our proposed method achieves significantly higher performance than the current SOTAs in CIFAR100, ImageNet-R, ImageNet-A, and CUB dataset. Our complexity analysis shows that our method requires a relatively smaller number of parameters and achieves moderate training time, inference time, and throughput. For further study, the source code of our method is available at https://github.com/anwarmaxsum/PROL.

This study proposes a challenging yet practical Federated Few-Shot Class-Incremental Learning (FFSCIL) problem, where clients only hold very few samples for new classes. We develop a novel Unified Optimized Prototype Prompt (UOPP) model to simultaneously handle catastrophic forgetting, over-fitting, and prototype bias in FFSCIL. UOPP utilizes task-wise prompt learning to mitigate task interference and over-fitting, unified static-dynamic prototypes to achieve a stability-plasticity balance, and adaptive dual heads for enhanced inferences. Dynamic prototypes represent new classes in the current few-shot task and are rectified to deal with prototype bias. Our comprehensive experimental results show that UOPP significantly outperforms state-of-the-art (SOTA) methods on three datasets with improvements up to 76% on average accuracy and 90% on harmonic mean accuracy respectively. Our extensive analysis shows UOPP robustness in various numbers of local clients and global rounds, low communication costs, and moderate running time. The source code of UOPP is publicly available at https://github.com/anwarmaxsum/FFSCIL.

The prompt-based approach has demonstrated its success for continual learning problems. However, it still suffers from catastrophic forgetting due to inter-task vector similarity and unfitted new components of previously learned tasks. On the other hand, the language-guided approach falls short of its full potential due to minimum utilized knowledge and participation in the prompt tuning process. To correct this problem, we propose a novel prompt-based structure and algorithm that incorporate 4 key concepts (1) language as input for prompt generation (2) task-wise generators (3) limiting matching descriptors search space via soft task-id prediction (4) generated prompt as auxiliary data. Our experimental analysis shows the superiority of our method to existing SOTAs in CIFAR100, ImageNet-R, and CUB datasets with significant margins i.e. up to 30% final average accuracy, 24% cumulative average accuracy, 8% final forgetting measure, and 7% cumulative forgetting measure. Our historical analysis confirms our method successfully maintains the stability-plasticity trade-off in every task. Our robustness analysis shows the proposed method consistently achieves high performances in various prompt lengths, layer depths, and number of generators per task compared to the SOTAs. We provide a comprehensive theoretical analysis, and complete numerical results in appendix sections. The method code is available in https://github.com/anwarmaxsum/LEAPGEN for further study.

The dependency on extensive expert knowledge for defining subgoals in hierarchical reinforcement learning (HRL) restricts the training efficiency and adaptability of HRL agents in complex, dynamic environments. Inspired by human-guided causal discovery skills, we proposed a novel method, Human Causal Perception and Inference-driven Hierarchical Reinforcement Learning (HCPI-HRL), designed to infer diverse, effective subgoal structures as intrinsic rewards and incorporate critical objects from dynamic environmental states using stable causal relationships. The HCPI-HRL method is supposed to guide an agent's exploration direction and promote the reuse of learned subgoal structures across different tasks. Our designed HCPI-HRL comprises two levels: the top level operates as a meta controller, assigning subgoals discovered based on human-driven causal critical object perception and causal structure inference; the bottom level employs the Proximal Policy Optimisation (PPO) algorithm to accomplish the assigned subgoals. Experiments conducted across discrete and continuous control environments demonstrated that HCPI-HRL outperforms benchmark methods such as hierarchical and adjacency PPO in terms of training efficiency, exploration capability, and transferability. Our research extends the potential of HRL methods incorporating human-guided causal modelling to infer the effective relationships across subgoals, enhancing the agent's capability to learn efficient policies in dynamic environments with sparse reward signals.

The existing approaches on continual learning (CL) call for a lot of samples in their training processes. Such approaches are impractical for many real-world problems having limited samples because of the overfitting problem. This article proposes a few-shot CL approach, termed flat-to-wide approach (FLOWER), where a flat-to-wide learning process finding the flat-wide minima is proposed to address the catastrophic forgetting (CF) problem. The issue of data scarcity is overcome with a data augmentation approach making use of a ball-generator concept to restrict the sampling space into the smallest enclosing ball. Our numerical studies demonstrate the advantage of FLOWER achieving significantly improved performances over prior arts notably in the small base tasks. For further study, source codes of FLOWER, competitor algorithms, and experimental logs are shared publicly in https://github.com/anwarmaxsum/FLOWER.

Federated Class Incremental Learning (FCIL) is a new direction in continual learning (CL) for addressing catastrophic forgetting and non-IID data distribution simultaneously. Existing FCIL methods call for high communication costs and exemplars from previous classes. We propose a novel rehearsal-free method for FCIL named prototypes-injected prompt (PIP) that involves 3 main ideas: a) prototype injection on prompt learning, b) prototype augmentation, and c) weighted Gaussian aggregation on the server side. Our experiment result shows that the proposed method outperforms the current state of the arts (SOTAs) with a significant improvement (up to 33%) in CIFAR100, MiniImageNet, and TinyImageNet datasets. Our extensive analysis demonstrates the robustness of PIP in different task sizes, and the advantage of requiring smaller participating local clients, and smaller global rounds. For further study, source codes of PIP, baseline, and experimental logs are shared publicly in https://github.com/anwarmaxsum/PIP.

Few-Shot Class-Incremental Learning (FSCIL)presents an extension of the Class Incremental Learning (CIL)problem where a model is faced with the problem of data scarcity while addressing the Catastrophic Forgetting (CF)problem. This problem remains an open problem because all recent works are built upon the Convolutional Neural Networks (CNNs)performing sub-optimally compared to the transformer approaches. Our paper presents Robust Transformer Approach (ROBUSTA)built upon the Compact Convolutional Transformer (CCT). The issue of overfitting due to few samples is overcome with the notion of the stochastic classifier, where the classifier's weights are sampled from a distribution with mean and variance vectors, thus increasing the likelihood of correct classifications, and the batch-norm layer to stabilize the training process. The issue of CFis dealt with the idea of delta parameters, small task-specific trainable parameters while keeping the backbone networks frozen. A non-parametric approach is developed to infer the delta parameters for the model's predictions. The prototype rectification approach is applied to avoid biased prototype calculations due to the issue of data scarcity. The advantage of ROBUSTAis demonstrated through a series of experiments in the benchmark problems where it is capable of outperforming prior arts with big margins without any data augmentation protocols.

Most few-shot learning works rely on the same domain assumption between the base and the target tasks, hindering their practical applications. This paper proposes an adaptive transformer network (ADAPTER), a simple but effective solution for cross-domain few-shot learning where there exist large domain shifts between the base task and the target task. ADAPTER is built upon the idea of bidirectional cross-attention to learn transferable features between the two domains. The proposed architecture is trained with DINO to produce diverse, and less biased features to avoid the supervision collapse problem. Furthermore, the label smoothing approach is proposed to improve the consistency and reliability of the predictions by also considering the predicted labels of the close samples in the embedding space. The performance of ADAPTER is rigorously evaluated in the BSCD-FSL benchmarks in which it outperforms prior arts with significant margins.

A continual learning (CL) model is desired for remote sensing (RS) image analysis because of varying camera parameters, spectral ranges, resolutions, etc. There exist some recent initiatives to develop CL techniques in this domain but they still depend on massive labeled samples which do not fully fit RS applications because ground truths are often obtained via field-based surveys. This article addresses this problem with a proposal of unsupervised flat-wide learning approach (UNISA) for unsupervised few-shot CL approaches of RS image scene classifications which do not depend on any labeled samples for its model updates. UNISA is developed from the idea of prototype scattering and positive sampling for learning representations while the catastrophic forgetting (CF) problem is tackled with the flat-wide learning approach combined with a ball generator to address the data scarcity problem. Our numerical study with RS image scene datasets and a hyperspectral dataset confirms the advantages of our solution. For future studies and reproductions, source codes of UNISA are shared publicly inhttps://github.com/anwarmaxsum/UNISA.

Low-cost transient resources such as Amazon's Elastic Compute Cloud (EC2) Spot instances can be opportunistically leveraged to reduce the ongoing costs of cloud applications. However, they are susceptible to unilateral revocations by the vendor making them a risky proposition for long-running applications with strict performance requirements. It is challenging to effectively balance the cost savings that transient resources provide with the associated revocation risk which, if realised, can impact application performance. To address this challenge, we propose an approach for risk-aware cloud cost optimization that is inspired by the concept ofportfolio diversification.Contract diversificationmitigates the revocation risk by procuring the required compute capacity as a mixed portfolio of transient and non-transient resources.Resource diversificationfurther diversifies the risk by using multiple transient resource types. Using our approach, consumers can leverage contract and resource diversification to proactively (re-)configure their application's resource portfolio to handle workload and resource price fluctuations while minimizing ongoing cost and keeping ongoing revocation risk within tolerable limits. Simulative evaluation using three real-world workload traces and Amazon's EC2 offerings demonstrate that our proposed approach can achieve meaningful cost savings compared to the baseline costs, while significantly reducing the portfolio's exposure to revocation risk.

In this paper, we introduce a new meta-model that captures core concepts for constructing software architectures for general-purpose, autonomous middleware frameworks that realize internalized and externalized self-adaptivity at both a system- and meta-level in order to achieve antifragility. The proposed meta-model builds on, specializes, and complements existing multi-agent meta-models in line with a previously published reference model for antifragile systems in the cyber domain.

Cloud consumers have access to an increasingly diverse range of resource and contract options, but lack appropriate resource scaling solutions that can exploit this to minimize the cost of their cloud-hosted applications. Traditional approaches tend to use homogeneous resources and horizontal scaling to handle workload fluctuations and do not leverage resource and contract heterogeneity to optimize cloud costs. In this paper, we propose a novel opportunistic resource scaling approach that exploits both resource and contract heterogeneity to achieve cost-effective resource allocations. We model resource allocation as anunbounded knapsack problem, and resource scaling as anone-step ahead resource allocation problem. Based on these models, we propose two scaling strategies: (a)delta capacity optimization, which focuses on optimizing costs for the difference between existing resource allocation and the required capacity based on the forecast workload, and (b)full capacity optimization, which focuses on optimizing costs for resource capacity corresponding to the forecast workload. We evaluate both strategies using two real world workload datasets, and compare them against three different scaling strategies. The results show that our proposed approach, particularly full capacity optimization, outperforms all of them and offers in excess of 70 percent cost savings compared to the traditional scaling approach.

In this paper, we study a large-scale heterogeneous task allocation problem in complex systems. Existing work on task allocation mainly tackles this well-known NP-hard problem from an optimisation perspective, where an exact or approximate solutions can be found after intensive computation. They have not been able to cater for the extra needs of scalability and robustness in large scale complex systems. In this work, we employ a game-theoretic framework to model the studied task allocation problem, and align the objective in task allocation (i.e., system optimality) with the concept of Nash equilibrium in game theory. Our formulation enables the expression of heterogeneity in both agents and tasks, and allows multiple agents to form teams or coalitions to cooperatively perform a task. Based on this formulation, we propose a novel GreedyNE algorithm to efficiently search for a Nash equilibrium solution. The proposed GreedyNE algorithm is a scalable, anytime, and monotonic algorithm, which in turn, makes it robust for the deployment in complex systems. GreedyNE is simple, easy to implement and flexible enough, so that it can also be used as a local search algorithm for improving the quality of any existing allocation solution. By conducting comprehensive experiments, we show that GreedyNE achieves a solution quality as good as the state-of-the-art approximation algorithms, yet with significantly lower computation time.

This paper explores the heterogeneous task allocation problem in Multi-robot systems. A game-theoretic formulation of the problem is proposed to align the goal of individual robots with the system objective. The concept of Nash equilibrium is applied to define a desired solution for the task allocation problem in which each robot can allocate itself to an appropriate task group. We also introduce a market-based distributed mechanism, called DisNE, to allow the robots to exchange messages with tasks and move between task groups, eventually reaching an equilibrium solution. We carry out comprehensive empirical studies to demonstrate that DisNE achieves near-optimal system utility in significantly shorter computation times when compared with the state-of-the-art mechanisms.

Cloud infrastructure providers are offering consumers a wide range of resource and contract options to choose from, yet most elasticity management solutions are incapable of leveraging this to optimize the cost and performance of cloudhosted applications. To address this problem, in this paper, we propose a novel resource scaling approach that exploits both resource and contract heterogeneity to achieve optimal resource allocations and better cost control. We model resource allocation as an Unbounded Knapsack Problem, and resource scaling as an one-step ahead resource allocation problem. Based on this, we present two scaling strategies, namely delta scale optimization and full scale optimization. Delta scale optimization supports the traditional notion of scaling resources horizontally, i.e., it computes an optimal allocation (or deallocation) of resources to increase (or decrease) the total compute capacity based on the current allocation and the forecast application workload. Full scale optimization, on the other hand, supports the notion of cost-optimal resource rescaling, i.e., the simultaneous allocation and deallocation of resources to meet the forecast workload irrespective of the decision to increase, decrease or maintain capacity. Both strategies provide users greater flexibility in managing trade offs between cost and performance. We motivate our research work by using a realistic and non-trivial scenario of resource scaling for a cloud-hosted IoT platform and use simple use cases to illustrate the benefit of our proposed approach.

Resilience is an important property of distributed cyber defence systems operating in complex, adversarial environments. While resilience can be realized through self-management, implementing distributed self-management poses several significant challenges. In this paper, we identify some of these challenges and present our initial work on an approach for cyber resilience called AWaRE. The novelty of AWaRE lies in the use of a conceptual, state-space-based design and reconstitution framework, combined with run-time models and distributed constraint satisfaction/optimization techniques for decision-making and coordination of system re-configurations. The run-time models are generated via an expressive domain-specific language enabling component, constraint and agent modeling as well as constraint problem decomposition and architectural self-organization. We realize AWaRE via a concrete software framework, focusing specifically on autonomic self-* properties pertaining to distributed system configuration, deployment and reliable operation. We demonstrate the value of AWaRE in the context of a simple, cloud-based enterprise scenario.

Consumers can realize significant cost savings by procuring resources from computational spot markets such as Amazon Elastic Compute Cloud (EC2) Spot Instances. They can take advantage of the price differentials across time slots, regions, and instance types to minimize the total cost of running their applications on the cloud. However, Spot markets are inherently volatile and dynamic, as a consequence of which Spot prices change continuously. As such, prospective bidders can benefit from intelligent insights into the Spot market dynamics that can help them make more informed bidding decisions. To enable this, we propose a descriptive statistics approach for the analysis of Amazon EC2 Spot markets to detect typical pricing patterns including the presence of seasonal components, extremes and trends. We use three statistical measures - the Gini coefficient, the Theil index, and the exponential weighted moving average. We also devise a model for estimating minimum bids such that the Spot instances will run for specified durations with a probability greater than a set value based on different look back periods. Experimental results show that our estimation yields on average a bidding strategy that can reliably secure an instance at least 80% of the time at minimum target guarantee between 50% and 95%.

Connected Communities (CCs) are socio-technical systems that rely on an information and communication technology (ICT) infrastructure to integrate people and organizations (companies, schools, hospitals, universities, local and national government agencies) willing to share information and perform joint decision-making to create sustainable and equitable work and living environments. We discuss a research agenda considering CCs from three distinct but complementary points of view: CC metaphors, models, and services.

The increasingly growing supply and demand for infrastructure as a service (IaaS) makes cloud trading possible in an open exchange (OCX) marketplace. The mechanisms based on economic principles show promise in addressing the problem of efficient cloud resource provisioning in such a marketplace, including resources allocation and pricing. Therefore, this article proposes effective and efficient economics-inspired double-sided mechanisms, focussing on attaining the allocative efficiency and truthfulness. Given non-deterministic polynomial time (NP) complexity of the considered problem and the computational tractability requirement for the practical solutions, we design and evaluate the approximation mechanisms for such markets. We propose a combinatorial greedy allocation mechanism to determine the distribution of cloud resources based on the sorted order of allocation candidates. We design the pricing mechanisms that derive the buyer prices based on critical-value, and the seller payments are determined via a mixed surplus-distribution rule that relies on direct and proportional-value payment. The theoretical analysis of the economic properties proves that the proposed mechanisms maintain Budget-Balance (BB), Individual Rationality (IR), Computational Tractability (CT), and achieve Truthfulness (T) for single-minded buyers. The experimental investigation of the approximation quality and the seller strategic manipulation reveal near-optimal allocation performance and near-truthful strategic incentive in our pricing mechanisms.

We propose a novel approach called, an orthogonal particle swarm optimization (OPSO) algorithm, for economic dispatch (ED) of thermal generating units (TGUs) in smart electric power gird (SEPG) environment. The characteristics of TGUs are nonlinear and the generation system becomes more and more complicated when these TGUs are subjected to ramp rate constraints and prohibited operating zones. In such case, the cost functions become non-smooth and non-convex due to the discontinuities in the cost curves. Moreover, for large-scale TGUs, the high dimensions used in ED problem become a big challenge to find global minimum and to avoid falling into local minima. The proposed OPSO algorithm has the ability to solve such complex problems including ED. The OPSO algorithm applies an orthogonal diagonalization process. It makes d particles (out of total m particles, m ≥ d) that have the possible solutions by constructing orthogonal vectors in the d-dimensional search space. These orthogonal vectors are generated and updated in each iteration and are utilized to guide the d particles to fly in one direction toward global minimum. The OPSO algorithm is evaluated and tested through 40 TGUs and its performance is compared with several other optimization methods. We found that the OPSO algorithm provides better results in term of cost under power grid constraints. Furthermore, we have shown that the OPSO algorithm significantly improves the PSO algorithm in terms of high solution quality, robustness and convergence.

Summary Cloud computing offers a realization of SOA in which IT resources are dynamically provisioned as services to consumers using flexible provisioning and pricing models. When provisioning such services, providers and consumers must first agree over the service usage terms and conditions, which are captured in Service Level Agreements (SLAs). In this paper, we propose a policy‐based framework with corresponding models, mechanisms and tools for the automated establishment of SLAs in open, diverse and dynamic cloud environments. The Automated SLA Management framework allows entities to specify their requirements and capabilities, and preferences over them in a flexible and expressive manner. It also supports multiple interaction models for SLA establishment, giving consumers and providers the flexibility to select the one that is most appropriate in a given context, while simultaneously participating in multiple concurrent SLA interactions using different interaction models. As part of the framework, we define a formal model for the underlying policies, a corresponding physical model WS‐SLAM that extends WS‐Policy and a reference architecture that can be easily implemented. We validate the practicability of our framework through the Smart CloudPurchaser prototype that can automatically purchase computing resources from Amazon EC2 under different scenarios and contexts. Copyright © 2013 John Wiley & Sons, Ltd.

This paper addresses the problem of flexible procurement of multiple distinct services characterized by multiple nonfunctional characteristics, i.e., quality-of-service attributes. We consider the one-to-many negotiation approach as a flexible method for procuring multiple different services by a buyer agent. We address the problem of coordinating the bidding strategy amongst multiple concurrent negotiations and propose novel dynamic negotiation strategies. The proposed strategies consider the behaviors of the opponents of the current negotiation encounter in managing the local reservation values of the common negotiation issues (attributes) of different services. Most previous works consider the problem of negotiation over a single object characterized by one or more issues. We extend our previous work and investigate a more complex situation where a buyer agent negotiates over multiple distinct services given that each service has multiple negotiation issues and multiple possible providers. The experimental results show evidence for the effectiveness and robustness of our dynamic negotiation strategies under various negotiation environments.

Migration to the cloud offers several benefits including reduced operational costs, flexibility, scalability, and a greater focus on business goals, but it also has a flip side reduced visibility. Organizations only have a blackbox view of cloud servers and while pricing and specification information is publicly available, there is limited information about cloud performance. This necessitates the need for tools that can provide greater visibility into cloud insfrastructure performance so that consumers can objectively compare and contrast the offerings from different providers. Smart CloudBench [1][2][3] is a system that allows users to run automated, ondemand, real-time and customized benchmark tests on cloud infrastructure. In this paper, we present Smart CloudMonitor - a performance monitoring tool that provides multi-layer performance monitoring capabilities to Smart CloudBench. It provides greater visibility and insight into cloud performance by monitoring both application performance as well as the corresponding resource consumption. Experiments conducted on cloud infrastructure using Smart CloudBench show the add value that Smart CloudMonitor provides to the process of cloud performance evaluation.

Cloud resource allocation and pricing is a significant and challenging problem for modern cloud providers, which needs to be addressed. In this work, we propose an adaptive greedy mechanism, which is a new type of greedy market mechanism for efficient cloud resource allocation. The mechanism is combinatorial and it is designed to be operated by a single cloud provider. We prove that our proposed market mechanism is truthful, i.e. the buyers do not have an incentive to lie about their true valuation for the resource. Our experimental investigation showed that the proposed mechanism outperforms the conventional (single-shot) approach for solving combinatorial auction in terms of generated social welfare and resource utilization.

Power balancing is an important issue when microgrids in island mode are considered.It requires active, real-time decision making to minimise the imbalances.El Farol Bar and Potluck problems are artificial problems designed to challenge the decision making process in a situation of limited information.They are theoretical, ill-defined problems designed to show that rational decision making in a situation of scarce information can give worse results than non-rational behavior.Potluck problem can be compared to electric power balancing in microgrids.But due to reduced complexity and constraints of theoretical problems and differences in goal functions the approach has to be in each case different.The study of these differences in this paper shows the importance of exchange of information between participants (or agents); it also suggest what type of information are necessary in what conditions.The amount of information that needs to be exchanged depends on the relation between participants (agents): the less cooperating their relation is, the less information they are willing to exchange.

As the rate of cloud computing adoption grows, so does the need for consumption assistance. Enterprises that are looking to migrate their IT systems to the cloud, would like to quickly identify providers that offer resources with the most appropriate pricing and performance levels to match their specific business needs. However, no two vendors offer the same resource configurations, pricing and provisioning models, making the task of selecting appropriate computing resources complex, time-consuming and expensive. In this paper, we present Smart CloudBench - a platform that automates the performance benchmarking of cloud infrastructure, helping potential consumers quickly identify the cloud providers that can deliver the most appropriate price/performance levels to meet their specific requirements. Users can estimate the actual performance of the different cloud platforms by testing representative benchmark applications under representative load conditions. Experimentation using the prototype implementation shows that higher price does not necessarily translate to better or more consistent performance, and benchmarking results can provide more information to help enterprises make better informed decisions.

As the rate of cloud computing adoption grows, so does the need for consumption assistance. Enterprises looking to migrate their IT systems to the cloud require assistance in identifying providers that offer resources with the most appropriate pricing and performance levels to match their specific business needs. In this paper, we present Smart Cloud Broker - a suite of software tools that allows cloud infrastructure consumers to evaluate and compare the performance of different Infrastructure as a Service (IaaS) offerings from competing cloud service providers, and consequently supports selection of the cloud configuration and provider with the specifications that best meet the user's requirements. Using Smart Cloud Broker, prospective cloud users can estimate the performance of the different cloud platforms by running live tests against representative benchmark applications under representative load conditions.

Automated negotiation is one of the most common approaches used to make decisions and manage disputes between computational entities leading them to optimal agreements. Many existing works tackle single-issue negotiations and the negotiation environment is assumed to be static so that the agents can make decisions based solely on the proposals of the counterparts and their own fixed parameters. Most real-world scenarios, however, involve complex domains and dynamic environments. In such cases, it is no longer sufficient to consider negotiation as an isolated activity in a static environment. Therefore, a more general framework for automated negotiation is needed in which the negotiation agents can be very flexible and adaptive. In this paper, we describe a generic framework for automated negotiation, which captures descriptively the social dynamics of the negotiation process. The proposed framework enables the agents to behave responsively to the changes in the environment. Their strategies can adapt as the conditions outside of the negotiation change to ensure that their decisions remain rational. And the agents are proactive and responsive by searching for options, which are outside of the negotiation and which may improve their outcomes. The key ideas and the overall system architecture together with a specific negotiation instance in a basic bilateral setting are described, along with two illustrative examples. The first example is in the context of e-commerce, and the second example is an application scenario of service level agreement negotiation in service computing. We also describe a prototypical implementation of the proposed negotiation framework.

Service-Oriented Computing (SOC) is emerging as a paradigm for developing distributed applications. A critical issue of utilizing SOC is to have a scalable, reliable, and robust service discovery mechanism. However, traditional service discovery methods using centralized registries can easily suffer from problems such as performance bottleneck and vulnerability to failures in large scalable service networks, thus functioning abnormally. To address these problems, this paper proposes a peer-to-peer-based decentralized service discovery approach named Chord4S. Chord4S utilizes the data distribution and lookup capabilities of the popular Chord to distribute and discover services in a decentralized manner. Data availability is further improved by distributing published descriptions of functionally equivalent services to different successor nodes that are organized into virtual segments in the Chord4S circle. Based on the service publication approach, Chord4S supports QoS-aware service discovery. Chord4S also supports service discovery with wildcard(s). In addition, the Chord routing protocol is extended to support efficient discovery of multiple services with a single query. This enables late negotiation of Service Level Agreements (SLAs) between service consumers and multiple candidate service providers. The experimental evaluation shows that Chord4S achieves higher data availability and provides efficient query with reasonable overhead.

We propose a policy-based framework for the automated establishment of SLAs for cloud computing services. The proposed framework supports multiple interaction models for SLA establishment giving consumers and providers the flexibility to choose one that is most appropriate in a given context, while simultaneously supporting multiple concurrent SLA interactions using different interaction models. We describe the underlying policies, focussing on the key features and contributions of the framework. We also validate our framework through a real-world use-case scenario using the Amazon EC2 service.

This paper addresses the problem of flexible procuring of multiple services with multiple non-functional characteristics, i.e., quality of service attributes. We investigate the one-to-many negotiation approach as a flexible method for procuring multiple services by a buyer agent. We address the problem of coordinating multiple concurrent negotiations and propose a novel dynamic negotiation strategy that considers the behaviors of the opponents in managing the local reservation values for the common negotiation issues. Most previous works consider the problem of negotiation over a single issue. We investigate a more complex scenario where a buyer agent negotiates with multiple seller agents over multiple services characterized by multiple issues. The initial experimental results show the effectiveness of our dynamic negotiation strategy when compared to a static strategy.

SUMMARY Cloud Computing is the latest paradigm proposed toward fulfilling the vision of computing being delivered as an utility such as phone, electricity, gas and water services. It enables users to have access to computing infrastructure, platform and software as services over the Internet. The services can be accessed on demand and from anywhere in the world in a quick and flexible manner, and charged for based on their usage, making the rapid and often unpredictable expansion demanded by nowadays' business environment affordable also for small spin‐off and start‐up companies. In order to be competitive, however, Cloud providers need to be able to adapt to the dynamic loads from users, not only optimizing the local usage and costs but also engaging into agreements with other Clouds so as to complement local capacity. The infrastructure in which competing Clouds are able to cooperate to maximize their benefits is called a Federated Cloud. Just as Clouds enable users to cope with unexpected demand loads, a Federated Cloud will enable individual Clouds to cope with unforeseen variations of demand. The definition of the mechanism to ensure mutual benefits for the individual Clouds composing the federation, however, is one of its main challenges. This paper proposes and investigates the application of market‐oriented mechanisms based on the General Equilibrium Theory of Microeconomics to coordinate the sharing of resources between the Clouds in the Federated Cloud. Copyright © 2010 John Wiley & Sons, Ltd.

SUMMARY The end‐to‐end QoS negotiation for service level agreement establishment for composite services involves compound multi‐party negotiations in which the composite service provider concurrently negotiates with multiple candidates for each atomic service, selecting the one that best satisfies the atomic service QoS preferences while ensuring that the end‐to‐end QoS requirements are also fulfilled. In order to be able to negotiate with potential candidates, it is necessary to derive the atomic utility boundaries from the global utility boundary. Additionally, there has to be a mechanism for updating these boundaries in subsequent negotiation rounds based on the individual negotiation outcomes. In this paper, we propose an algorithm for the decomposition of global utility boundary into atomic service utility boundaries, and the surplus redistribution from successful negotiation outcomes among the remaining negotiations. The proposed mechanism is a practical approach to efficiently coordinate concurrent service negotiations within complex workflows, enabling the iterative and interactive adjustment of the negotiation boundaries for each atomic service in a composition based on the performance of other atomic negotiations. We demonstrate the feasibility of our approach by evaluating it with some popular negotiation strategies using the Specialized Property Search Scenario. Copyright © 2011 John Wiley & Sons, Ltd.

We propose a policy-based framework for the QoS differentiated provisioning of services. The proposed frame-work improves the state-of-the-art in policy-based preference specification by combining cardinal and ordinal preferences. We describe the underlying models, focussing on the key features and contributions of the proposed framework. We also show how, using our framework, the QoS evaluation problem can be translated to a Constraint Satisfaction Problem while preserving the semantics of the preference policies.

We present a policy-centered QoS meta-model which can be used by service providers and consumers alike to express capabilities, requirements, constraints, and general management characteristics relevant for SLA establishment in service aggregations. We also provide a QoS assertion model which is generic, domain-independent and conforming to the WS-Policy syntax and semantics. Using these two models, assertions over acceptable and required values for QoS properties can be expressed across the different service layers and service roles.

Automated bilateral agent negotiation under uncertainty, that is with imprecise or uncertain information about preferences, utilities and strategies of the opponent is known to be hard. In this paper, we present the first adaptive solution that bases on multistage fuzzy decision making. The modelling of individual preferences as fuzzy goal and fuzzy constraints, and observed strategic concession behaviour of opponents during negotiation as a fuzzy Markov decision process allows the agent to adapt its negotiation strategies and implied behaviour to improve its individual payoffs. In particular, we show that such adaptive bilateral negotiation strategies can be efficiently derived by an agent from negotiation threads of only two reference cases in its respectively maintained fuzzy transition matrix. Finally, we demonstrate the benefit of applying this solution to different soft-constrained negotiation settings by an initial comparative experimental evaluation.

The development of mechanisms to understand and model the expected behaviour of multiagent learners is becoming increasingly important as the area rapidly find application in a variety of domains. In this paper we present a framework to model the behaviour of Q-learning agents using the o-greedy exploration mechanism. For this, we analyse a continuous-time version of the Q-learning update rule and study how the presence of other agents and the o-greedy mechanism affect it. We then model the problem as a system of difference equations which is used to theoretically analyse the expected behaviour of the agents. The applicability of the framework is tested through experiments in typical games selected from the literature.

The assurance of quality-of-service (QoS) is critical for the successful deployment of service-oriented applications, especially in open, dynamic, and distributed cross-organizational environments. Adaptive management of the QoS assured provision of composite services is required for more reliable, fault-tolerant, and flexible service delivery in such environments. It can be realized with software agents offering a unified framework and necessary capabilities for carrying out different adaptive management tasks across the whole lifecycle of composite service provision.

Market-based mechanisms offer a promising approach for distributed resource allocation. In this paper we consider the Iterative Price Adjustment, a pricing mechanism that can be used in commodity-market resource allocation systems. We address the sce

In the Web services environment, service level agreements (SLA) refer to mutually agreed understandings and expectations of service provision between service consumers and providers. Although management of SLA is critical to wide adoption of Web services technologies in the real world, support for it is very limited nowadays. There lacks adequate frameworks and technologies supporting various SLA operations. This paper presents an agent-based framework which utilises the agents' ability of negotiation, interaction, and cooperation to facilitate autonomous and flexible SLA management. Based on this framework, mechanisms for autonomous SLA formation, recovery, and profiling are proposed and discussed.

This paper reports innovative research aiming at supporting autonomous establishment and maintenance of service level agreements in order to guarantee end-to-end quality of service requirements for service composition provision. In this research, a set of interrelated service level agreements is established and maintained for a service composition, through autonomous agent negotiation. To enable this, an innovative framework is proposed in which agents on behalf of the service requestor and the service providers can negotiate service level agreements in a coordinated way. This framework also enables adaptive service level agreement re-negotiation in the dynamic and ever-changing service environment

Web service compositions have attracted considerable efforts in the context of supporting enterprise application integrations. For a composite service, in addition to its functional requirements, QoS requirements are important and deserve a special attention. The central question to a QoS composition problem is how to compose a service from different subcomponent services so that its overall QoS can satisfy certain requirements. In this paper, we propose an agent-based method using fuzzy distributed constraint satisfaction problem (fuzzy DisCSP) techniques to solve this problem. We show that by using the composition structures, local constraints can be constructed and used with DisCSP. We also present an a new algorithm called the fuzzy constraint satisfaction algorithm for distributed environment (FADE) to solve the problem and discuss our experiment in building a prototypical system to prove the feasibility of our approach

The ASAPM project aims at developing new techniques, mechanisms and software solutions for enablement of flexible, dynamic and robust management of serviceoriented application provision processes to ensure collective functionality, end-to-end QoS and stateful coordination of complex services.

Presents an experimental toolkit for modeling and solving fuzzy constraint satisfaction problems (FCSPs) called JFSolver. JFSolver is a Java based library developed as an extension of a crisp constraint programming library to provide an enhanced functionality for handling fuzzy constraints including fuzzy constraint specification, propagation and search mechanisms. JFSolver comprises an application programming interface for fuzzy constraint programming and a graphical editor for interactive problem definition and solving. This functionality is demonstrated with an illustrative example.

This paper overviews an experimental fuzzy e-negotiation agents system, FeNAs, that can support automated negotiation in the presence of imprecise information. The system uses the principles of fuzzy constraint-based reasoning involving fuzzy constraint modeling, satisfaction and propagation. It is demonstrated with a prototype for the used car-trading problem. The system supports multi-issue negotiations where offers consist of a number of issues that can include the price of the car and other value-added services such as warranty and the value of the trade-in car. The agents exchange offers on the basis of the information available and negotiation strategies used by each party. Information available to both the buyer and the seller can include the make, model, color, transmission, age and mileage of the car. Each agent has also some private information including preferences, priorities and financial constraints that are not available to other agents. This information can be imprecise where constraints, preferences and priorities are defined as fuzzy constraints describing the level of satisfaction of an agent (and its user) with different potential solutions. The overall objective of an agent is to find a solution that maximizes the agent's utility at the highest possible level of constraint satisfaction subject to its acceptability by other agents. During negotiation the agents follow a common protocol of negotiation and individual negotiation strategies.