Automated decision support might be insufficient for managing wicked problems; instead, new social arrangements might be required to alter the trajectory of a social system. Since socio-technical systems will comprise human and computational intelligences interacting, we can leverage that interaction by using Artificial Intelligence (AI) to build innovation-support systems (ISS), which interleaves human expertise, generative AI and multi-agent simulation to imagine alternative social arrangements and envision societal trajectories. We define a socially-guided machine-learning framework that provides a ‘safe space’ for exploration and an evidence-base for selecting and modifying social arrangements. The significance of socially guided machine learning for innovation support is to assist societies with mitigating wicked problems through co-production with AI.

Polycentricity theory is predicated on the existence of multiple decision-making centres within a metropolitan area, collectively using a mutually agreed conventional system of governance. The digital transformation has increasingly integrated agentic computational artefacts (or artificial intelligence systems) into these social arrangements, thereby extending the concept of polycentricity to encompass ‘digital polycentricity.’ Digital polycentricity can be examined through four distinct ‘lenses’: interactional, governmental, architectural, and axiological. In this paper, we employ the interactional lens to study artificial interaction. Our analysis highlights several critical issues related to power and labour in the digital society. To mitigate these issues, we propose two strategies for developing Value-Sensitive Operationalisation (VSO), a framework aimed at guiding the creation of socio-technical systems that are fit-for-purpose and contextually appropriate. These strategies include: 1) leveraging digital data objects to support collective action, and 2) embedding a sense of place in the design of algorithms. By using these strategies, VSO practitioners can develop artificial intelligence (AI) systems that effectively meet user needs and cultivate a human–AI relationship designed to re-empower human stakeholders.

Many large-scale platforms for social coordination suffer from a lack of reuse and opportunities for domain-specific customisation, yet offer a soft-target single point of failure for security exploits. Moreover, they are mostly privately-owned by trans-national corporations, whose motives and values are sometimes distinctly different from their user communities, wherein Artificial Intelligence (AI) is often used as a means to abstract or reinforce pre-existing asymmetric power relationships and inequalities. To address these problems, this paper presents the development of a system comprising a platform (an extensible architecture for social coordination) and meta-platform (a system for instantiating the platforms), which maps a high-level architecture, design methodology and operationalisation model to a platform instantiation, according to five design “values” of data autonomy, sustainability, usability, customisability, and multiplicity. Each of these values can then be cross-referenced with three further supra-functional requirements - platform democratisation, knowledge codification and continuous self-improvement - which we argue offers a new approach to platformisation using the lens of ethical platformisation.

The global challenges posed by the simultaneous de-mographic transitions of urbanisation and ageing have prompted the World Health Organisation to develop the Age-Friendly City (AFC) approach. This approach, however, is overly reliant on spatial solutions to effectively respond to a wicked problem which has no obvious terminating condition, requires continuous adaption as needs, values and contributions change, and is highly dependent on an equitable distribution of resources. To address this, the paper explores the development of a methodology for a socio-technical system (STS) that facilitates the creation of an AFC in a borough of Central London. However, it is concluded that such a methodology cannot yet be operationalised and requires the expansion of STS theory to include a spatial dimension, the development of an effective participatory design methodology, and a focus on contributive justice to ensure meaningful participation.

In many social, cyber-physical, and socio-technical systems, a group of autonomous peers can encounter a knowledge aggregation problem, requiring them to organise themselves, without a centralised authority, as a distributed information processing unit (DIP). In this article, we specify and implement a new algorithm for knowledge aggregation based on Nowak's psychological theory Regulatory Theory of Social Influence (RTSI). This theory posits that social influence consists of not only sources trying to influence targets, but also targets seeking sources by whom to be influenced and learning what processing rules those sources are using. A multi-agent simulator SMARTSIS is implemented to evaluate the algorithm, using as its base scenario a linear public goods game where the DIP's decision is a qualitative question of distributive justice. In a series of experiments examining the emergence of expertise, we show how RTSI enhances the effectiveness of the multi-agent DIP as a social group while conserving each agent's individual resources. Additionally, we identify eight criteria for evaluating the DIP unit's performance, consisting of four conflicting pairs of systemic drivers, and discuss how RTSI maintains a balanced tension between the four driver pairs through the emergence and divergence of expertise. We conclude by arguing that this shows how psychological theories like RTSI can have a crucial role in informing agent-based models of human behaviour, which in turn may be critically important for effective knowledge management and reflective self-improvement in both cyber-physical and socio-technical systems.

Cooperative survival "games" are situations in which, during a sequence of catastrophic events, no one survives unless everyone survives. Such situations can be further exacerbated by uncertainty over the timing and scale of the recurring catastrophes, while the resource management required for survival may depend on several interdependent subgames of resource extraction, distribution, and investment with conflicting priorities and preferences between survivors. In social systems, self-organization has been a critical feature of sustainability and survival; therefore, in this article we use the lens of artificial societies to investigate the effectiveness of socially constructed self-organization for cooperative survival games. We imagine a cooperative survival scenario with four parameters: scale, that is, n in an n-player game; uncertainty, with regard to the occurrence and magnitude of each catastrophe; complexity, concerning the number of subgames to be simultaneously "solved"; and opportunity, with respect to the number of self-organizing mechanisms available to the players. We design and implement a multiagent system for a situation composed of three entangled subgames-a stag hunt game, a common-pool resource management problem, and a collective risk dilemma-and specify algorithms for three self-organizing mechanisms for governance, trading, and forecasting. A series of experiments shows, as perhaps expected, a threshold for a critical mass of survivors and also that increasing dimensions of uncertainty and complexity require increasing opportunity for self-organization. Perhaps less expected are the ways in which self-organizing mechanisms may interact in pernicious but also self-reinforcing ways, highlighting the need for some reflection as a process in collective self-governance for cooperative survival.

The increase in computational and communication power of pervasive devices is also enabling to embed increasing intelligence in devices, there included the capability to act in autonomy, and possibly interacting with each other, in order to achieve specific goals, thus leaving humans out of the decision loop. In this context, the goal of the panel is thus reasoning about the possible implications (technical, ethical, and lega) of assigning great and often critical decision power to pervasive autonomous systems.

This short paper reports the summary of the panel held in the context of the PerCom 2022 conference.

As artificial intelligence and smart devices increasingly infiltrate everyday life, cooperative interactions between humans and computers are correspondingly becoming more common. Errors are an inevitability in these interactions and can destabilise long-term working relationships. In this work, three online games of increasing difficulty (N=2037) were designed where participants played a cooperative game with an artificial AI player and encountered an unexpected error. Different training methods were used to establish a rapport between the human and AI players. Overall, despite answering correctly, participants were increasingly more likely to say they had made a mistake and that they were to blame as the difficulty of the game increased. Since participants were also unaware of the extent of their exposure to AI, this study shows that there is a tendency to apply preconceived expectations of AI and misattribute error origination which, if not addressed, could lead to critical breakdowns of trust.

The diversity between individual learners in an ensemble is known to influence its performance. However, there is no standard agreement on how diversity should be defined, and thus how to exploit it to construct a high-performing classifier. We propose two new behavioural diversity metrics based on the divergence of errors between models. Following a neuroevolution approach, these metrics are then used to guide a novelty search algorithm to search a space of neural architectures and discover behaviourally diverse classifiers, iteratively adding the models with high diversity score to an ensemble. The parameters of each ANN are tuned individually with a standard gradient descent procedure. We test our approach on three benchmark datasets from Computer Vision --- CIFAR-10, CIFAR-100, and SVHN --- and find that the ensembles generated significantly outperform ensembles created without explicitly searching for diversity and that the error diversity metrics we propose lead to better results than others in the literature. We conclude that our empirical results signpost an improved approach to promoting diversity in ensemble learning, identifying what sort of diversity is most relevant and proposing an algorithm that explicitly searches for it without selecting for accuracy.

A growing sense of unfairness permeates our quasi-digital society. Despite drivers supporting and motivating ethical practice in the digital technology ecosystem, there are compounding barriers to fairness that, at every level, impact technology innovation, delivery and access. Amongst these are barriers and omissions at the earliest stages of technology intentionality and design; systemic inadequacies in sensing systems that deteriorate performance for individuals based on ethnicity, age and physicality; system design, co-requisite and interface decisions that limit access; biases and inequities in datasets and algorithms; and limiting factors in system function and security. Additionally, there are concerns about unethical and illegal practices amongst digital technology providers: for example, in planned obsolescence and anti-competitive behaviors, failings in data practices and security, and in responses to problematic use and behaviors. It is critical that these failings are identified and addressed to better evolve a fairer future digital technology ecosystem. This paper contributes a perspective on technological stewardship and innovation; it identifies the compounding nature of barriers to fairness in the current digital technology ecosystem, and contrasts these with the non-compounding fairness drivers that, in general, establish minimum requirements.

Community energy systems (CESs) are shared energy systems in which multiple communities generate and consume energy from renewable resources. At regular time intervals, each participating community decides whether to self-supply, store, trade, or sell their energy to others in the scheme or back to the grid according to a predefined policy which all participants abide by. The objective of the policy is to maximise average satisfaction across the entire CES while minimising the number of unsatisfied participants. We propose a multi-class, multi-tree genetic programming approach to evolve a set of specialist policies that are applicable to specific conditions, relating to abundance of energy, asymmetry of generation, and system volatility. Results show that the evolved policies significantly outperform a default handcrafted policy. Additionally, we evolve a generalist policy and compare its performance to specialist ones, finding that the best generalist policy can equal the performance of specialists in many scenarios. We claim that our approach can be generalised to any multi-agent system solving a common-pool resource allocation problem that requires the design of a suitable operating policy.

The so-called fourth industrial revolution and its economic and societal implications are no longer solely an academic concern, but a matter for political as well as public debate. Characterized as the convergence of robotics, AI, autonomous systems and information technology – or cyberphysical systems – the fourth industrial revolution was the focus of the World Economic Forum, at Davos, in 2016[1]. Also in 2016 the US White House initiated a series of public workshops on artificial intelligence (AI) and the creation of an interagency working group, and the European Parliament Committee for Legal Affairs published a draft report with recommendations to the Commission on Civil Law Rules on Robotics.

Many instances of socio-technical systems in the digital society and digital economy require some form of self-governance. Examples include community energy systems, peer production systems, participatory sensing applications, and shared management of communal living areas or workspace. Such systems have several features in common, of which three are that they are rule-oriented , self-organising, and value-sensitive , and in operation, this combination of features entails self-modification of the rules in order to satisfice a changeable set of values. This presents a fundamental dilemma for systems design. On the one hand, the system must be sufficiently unrestricted (resilient, flexible) to enable a diverse group but with a shared set of congruent values to achieve their joint purposes in collective action situations. On the other hand, it must be sufficiently restricted (stable, robust) to prevent a subset of the group from exploiting self-determination ‘against itself’ and usurp control of the system for the benefit of its own narrow interests. To address this problem, we consider a study of classical Athenian democracy which investigates how the governance model of the city-state flourished. The work suggests that exceptional knowledge management , i.e., making information available for socially productive purposes, played a crucial role in sustaining its democracy for nearly 200 years, by creating processes for aggregation, alignment, and codification of knowledge. We therefore examine the proposition that some properties can be generalised to resolve the rule-restriction dilemma by establishing a set of design principles intended to make knowledge management processes open, inclusive, transparent, and effective in self-governed social technical systems. We operationalise three of these principles in the context of a collective action situation, namely self-organised common-pool resource allocation, and present the results of a series of experiments showing how knowledge management processes can be used to obtain robust solutions for the perception of fairness, allocation decision, and punishment mechanisms. By applying this operationalisation of the design principles for knowledge management processes as a complement to institutional approaches to governance, we demonstrate empirically how it can satisfice shared values, distribute power fairly, and apply “common sense” in dealing with rule violations. We conclude by arguing that this approach to the design of socio-technical systems can provide a balance between restricted and unrestricted self-modification of conventional rules, and can thus provide the foundations for sustainable and democratic self-governance in socio-technical systems.

It is a recurring requirement in open systems, such as networks, distributed systems, and socio-technical systems, that a group of agents must coordinate their behaviour for the common good. In those systems—where agents are heterogeneous—unexpected behaviour can occur due to errors or malice. Agents whose practices free-ride the system can be accepted to a certain level; however, not only do they put the stability of the system at risk, but they also compromise the agents that behave according to the system’s rules. In social systems, it has been observed thatsocial capitalis an attribute of individuals that enhances their ability to solve collective action problems. Sociologists have studied collective action through human societies and observed that social capital plays an important role in maintaining communities though time as well as in simplifying the decision-making in them. In this work, we explore the use of Electronic Social Capital for optimising self-organised collective action. We developed a context-independent Electronic Social Capital framework to test this hypothesis. The framework comprises a set of handlers that capture events from the system and update three different forms of social capital: trustworthiness, networks, and institutions. Later, a set of metrics are generated by the forms of social capital and used for decision-making. The framework was tested in different scenarios such as two-player games,n-player games, and public goods games. The experimental results show that social capital optimises the outcomes (in terms of long-term satisfaction and utility), reduces the complexity of decision-making, and scales with the size of the population. This work proposes an alternative solution using Electronic Social Capital to represent and reason with qualitative, instead of traditional quantitative, values. This solution could be embedded into socio-technical systems to incentivise collective action without commodifying the resources or actions in the system.

In recent years, the significant advancements in miniaturised computing and pervasive communication networks have paved the way for ubiquitous computing environments. In such environments users interact with systems through novel and implicit methods. In this context, affective computing provides a dimension of interaction, raising a number of opportunities to address not only utilitarian but also hedonic needs. At the same time, a number of challenges arise beyond the technical aspects, that are related to the individual and other societal implications. A review of the main opportunities and challenges is presented, supporting the identification of the main requirements for the design and development of systems in interactive pervasive environments, focusing on an affective loop of interaction. A framework is proposed, identifying main modules and functionality alongside a methodology to instantiate in specific domains of application.

IEEE SASO (Self-Adapting and Self-Organizing Systems) is the premier international conference for computer systems and networks that autonomously change some aspect of themselves: code, form, function, shape, structure, components, and so on. Over the past 10 years, it has emerged as a key multidisciplinary event for sharing theoretical insights and technical innovations across the numerous scientific fields and application domains impacted by this research. In 2013, the conference was hosted by Drexel University in Philadelphia, Pennsylvania, and offered an extensive program of high-quality papers. This special issue of ACM TAAS presents selected, revised, and extended best articles, which showcase the rich variety and depth of the SASO scientific community.

In this article, we complement Elinor Ostrom’s institutional design principles for enduring common-pool resource management with Nicholas Rescher’s theory of distributive justice based on the canon of legitimate claims. Two of Ostrom’s principles are that the resource allocation method should be congruent with the local environment, and that those affected by the allocation method (the appropriators) should participate in its selection. However, these principles do not say anything explicitly about the fairness of the allocation method or the outcomes it produces: for this, we need a mechanism for distributive justice. Rescher identified a number of different mechanisms, each of which had both its merits and demerits, and instead maintained that distributive justice consisted in identifying the legitimate claims in context, accommodating multiple claims in case of plurality, and reconciling them in case of conflict. Accordingly, we specify a logical axiomatisation of the principles with the canon of legitimate claims, whereby a set of claims is each represented as a voting function, which collectively determine the rank order in which resources are allocated. The appropriators vote on the weight attached to the scoring functions, and so self-organise the allocation method, taking into account both the plurality of and conflict between the claims. Therefore, the appropriators exercise collective choice over the method, and the method itself is congruent with the local environment, taking into account both the resources available and the relative claims of the appropriators. Experiments with a variant of the linear public good game show that this pluralistic self-organising approach produces a better balance of utility and fairness (for agents that comply with the rules of the game) compared to monistic or fixed approaches, provide “fairness over time” (a series of ostensibly unfair individual allocations is revealed to be cumulatively fair), and offer an intuition of how to resolve the free-rider phenomenon in provision and appropriation of common-pool resources.

We propose a model of vicarious reinforcement in rule-based learning agents. The influence of this reinforcement is investigated in a population where a law is enforced ex ante. The norm-governed population of learning agents is formalised and simulated in an executable probabilistic rule-based argumentation framework. Vicarious experiences are expressed with rules and their learning effects are integrated into reinforcement learning. So, agents learn not only from their own experiences but also by taking into account the experiences of others. We show that simulation results differ from traditional calculus based on expected utilities.

In recent years, the technological developments in mobile and communication networks have paved the way for smart environments, whose final goal is to provide users with enhanced experiences. The measure of user experience satisfaction, or quality of experience, may be defined as an affective state in response to a service. Thus, an experiment was devised to explore the relationship between users' affective state and their context, for assessing quality of experience in urban public transport services. A pilot study, conducted to evaluate the feasibility and requirements of such an experiment is presented, leading to a large scale field study.

The convergence of personal devices, pervasive communication networks and remote computing has caused a fundamental shift in the user interaction paradigm. Multiple methods have enabled an implicit loop of interaction that goes beyond the traditional graphical interfaces. Human emotion is one of such dimensions, supporting the development of empathic systems. Thus, quality of user experience, a subjective measure, may be defined as the resulting affective state from an interaction, which can be dynamically assessed. In mobile ubiquitous settings, leveraging this affective interaction for providing personalisation and immersive digital services has the potential to significantly impact user experience. This paper investigates the relationship between user affect and experience in the context of urban public transport.

Open computing systems, from sensor networks to Smart-Grids, face the same challenge: a set of autonomous, heterogenous agents, needing to collectivise and distribute resources without a centralised decision-making authority. One possible solution is to collectively agree a set of rules for resource provision and appropriation, but there remains the possibility for some agents not to comply with the rules. Therefore, some notion of “trust” with respect to rule compliance is essential. In this talk, we use a notion of trust stemming from analytic philosophy, whereby “trust” is analysed as a belief that there is a rule and an expectation that someone else's behaviour will conform to the rule, underpinned by a reparation mechanism if it does not. In the context of an electronic institution, whose rules are based on Elinor's Ostrom's principles of enduring institutions for common-pool resource management, we consider how this notion of “trust” can be represented by principles of distributive and retributive justice, and we report some experimental results in self-organisation of the “justice system”.

We address the problem of engineering self-organizing electronic institutions for resource allocation in open, embedded, and resource-constrained systems. In such systems, there is decentralized control, competition for resources and an expectation of both intentional and unintentional errors. The “optimal” distribution of resources is then less important than the endurance of the distribution mechanism. Under these circumstances, we propose to model resource allocation as a common-pool resource management problem, and develop a formal characterization of Elinor Ostrom’s socio-economic principles for self-governing institutions. This article applies a method for sociologically inspired computing to give a complete axiomatization of six of Ostrom’s eight principles in the Event Calculus. A testbed is implemented for experimenting with the axiomatization. The experimental results show that these principles support enduring institutions, in terms of longevity and membership, and also provide insight into calibrating the transaction and running costs associated with implementing the principles against the behavioral profile of the institutional membership. We conclude that it is possible to express Ostrom’s principles in logical form and that they are necessary and sufficient conditions for enduring self-organizing electronic institutions to manage sustainable common-pool resources.

Electronic markets, dispute resolution and negotiation protocols are three types of application domains that can be viewed as open agent societies. Key characteristics of such societies are agent heterogeneity, conflicting individual goals and unpredictable behavior. Members of such societies may fail to, or even choose not to, conform to the norms governing their interactions. It has been argued that systems of this type should have a formal, declarative, verifiable, and meaningful semantics. We present a theoretical and computational framework being developed for the executable specification of open agent societies. We adopt an external perspective and view societies as instances of normative systems. In this article, we demonstrate how the framework can be applied to specifying and executing a contract-net protocol. The specification is formalized in two action languages, the C + language and the Event Calculus, and executed using respective software implementations, the Causal Calculator and the Society Visualizer. We evaluate our executable specification in the light of the presented case study, discussing the strengths and weaknesses of the employed action languages for the specification of open agent societies.

The computers at cyber centres are vulnerable. They can be used for various illegal activities. Only the system-level security solutions are not sufficient to prevent these activities completely. There is a need for an intelligent scheme which identifies culprits based on behaviours and actions. This paper proposes an intelligent and environment based proactive security system for cyber centres, using cognitive agents. The agents are used for belief generation, belief analysis and to execute suitable security actions and pro-actions. This method of controlling the cyber crime at usage level, will be the requisite step towards preventing Internet abuse.

Three studies investigated whether users' strategies for customising online avatars increase their self-focused attention, also known as private self-awareness. Study 1 showed that a high number of users adapt their avatars toreflect their own appearance. Study 2 demonstrated that users who perceive their avatars to be similar to their own appearance experience as a result heightened private self-awareness. In Study 3, private self-awareness pervadedsocial interaction taking place over time when users with representative avatars, compared to a control group, reported increased private self-awareness. Drawing from research in interpersonal communication, we suggest that avatars which increase their owners' self-focus may have an influence on online behavior in the context of social computing.

Voting is an essential element of mechanism design for multi-agent systems, and by extension applications built on such systems, which includes ad hoc networks, virtual organizations, and decision support tools implementing online deliberative assemblies. Much attention has been given both to designing the process so that it is resistant to manipulation by strategic voting and so that an automated system can follow rules of order as developed for the conduct of formal meetings. In this paper, we define, characterize, formally specify and animate a general voting protocol. In particular we show how the requirements established by the characterization are captured by the specification and are exhibited by the animation. The importance of these requirements is in ensuring robustness by respecting the way in which votes are cast and the outcome is declared, especially as this issue relates to the 2004 ACM Statement on E-Voting.

Voting is an essential element of mechanism design for multi-agent systems, and decision support for CSCW tools implementing online deliberative assemblies. Much attention has been given both to designing the process so that it is resistant to manipulation by strategic voting, and so that an automated system can follow rules of order as developed for the conduct of formal meetings. In this paper, we formalise a general voting protocol trying to take into account a right to cast a vote, and an entitlement that the vote cast is counted in the correct way. We discuss the design and development of a system for online deliberative assemblies, that incorporates this protocol as part of a suite of protocols which collectively implement rules of order. We conclude with some comments on the voting protocol as it relates to the 2004 ACM Statement on E-Voting.

This paper describes the specification and implementation of a new three-layer time-aware agent architecture. This architecture is designed for applications and environments where societies of humans and agents play equally active roles, but interact and operate in completely different time frames. The architecture consists of three layers: the April real-time run-time (ART) layer, the time aware layer (TAL), and the application agents layer (AAL). The ART layer forms the underlying real-time agent platform. An original online, real-time, dynamic priority-based scheduling algorithm is described for scheduling the computation time of agent processes, and it is shown that the algorithm's O(n) complexity and scalable performance are sufficient for application in real-time domains. The TAL layer forms an abstraction layer through which human and agent interactions are temporally unified, that is, handled in a common way irrespective of their temporal representation and scale. A novel O(n2) interaction scheduling algorithm is described for predicting and guaranteeing interactions' initiation and completion times. The time-aware predicting component of a workflow management system is also presented as an instance of the AAL layer. The described time-aware architecture addresses two key challenges in enabling agents to be effectively configured and applied in environments where humans and agents play equally active roles. It provides flexibility and adaptability in its real-time mechanisms while placing them under direct agent control, and it temporally unifies human and agent interactions.

Open multi-agent computational systems are composed of heterogeneous and possibly antagonistic software entities. Characteristic features are limited trust and unpredictable behaviour. Members of such systems may fail to, or even choose not to, conform to the norms governing their interactions. It has been argued that systems of this type should have a formal, declarative, verifiable, and meaningful semantics. We present a theoretical and computational framework being developed for the executable specification of such systems. We adopt an external perspective and view open computational systems as instances of normative systems. In this paper we demonstrate how the framework can be applied to specifying and executing an argumentation protocol based on Brewka's reconstruction of Rescher's theory of formal disputation. The specification is formalised in the action language C+ and executed using the 'Causal Calculator' (CCALC) implementation.

This paper reports on and discusses recent research into software agent mediated retailing and considers how it may act as a catalyst to the formation of local electronic retail communities. Each member of the e-retail community has their own software agent that acts on their behalf. Their software agent acts as an electronic personal sales assistant (ePSA), interacting with the person and assisting with the selection and purchase of products. We consider issues that arise when each person can express their likes and dislikes for various aspects of the product range to their software agent, both in terms of the software agent's ability to serve them better, and how such preferences may be shared with others in the larger community to help them—and others in the community—make better and more informed choices. The investigations reported use the multimedia access through personal persistent agents (MAPPA) system, an experimental kiosk based e-retail system. MAPPA integrates a software agent based architecture, a novel form of product display and the use of an animated character agent to enhance the user's sense of personalisation. We report on user evaluation studies of the MAPPA system to investigate the effectiveness of the character agent as an interface to the e-retail system using both conventional usability evaluation techniques and eye-tracking technology. We also describe a simulation program that allows us to investigate the dynamics of e-retail community cohesion when a number of different parameters are considered, including the characteristics of the individual community members and the algorithms employed by the software agents to share information. Lastly, we develop and discuss notions of loyalty, trust, reputation and preference, four issues critical to the development and maintenance of the relationships between individual people, their personal software agents, businesses and the larger e-retail connected community. We propose these issues as ‘supra-functional’ criteria for human–computer interaction design in this area.

Efficient allocation of services to form a supply chain to solve complex tasks is a crucial problem. Optimal service allocation based on a single criterion is NP-Complete. Furthermore, complex tasks in general have multiple criteria that may be conflicting and non-commensurable. This paper presents a two-stage brokering algorithm for optimal anytime service allocation based on multiple criteria. In the first stage, a hierarchical task network planner is used to identify the services required to solve a task. In the second stage, a genetic algorithm (GA) determines service providers based on multiple criteria to provide the services identified by the planner. We present our algorithm and results from various experiments conducted to analyze the effect of various parameters that influence the complexity of the problem. In general, the results show the GA finds optimal solutions much quicker than a standard search algorithm. The empirical results also indicate the performance of the algorithm is sub-linear or polynomial time for various parameters. The algorithm has the ability to deal with any number of criteria. By addressing this problem, we expand the range of problems being addressed to any that require simultaneous optimization of multiple criteria and/or planning.