Agents Infrastructure Applications and Norms Michael Luck University

Agents, Infrastructure, Applications and Norms Michael Luck University of Southampton, UK

Overview § Monday • Agents for next generation computing § Agent. Link Roadmap § Tuesday • The case for agents • Agent Infrastructure § Conceptual: SMART § Technical: Paradigma/act. SMART • Agents and Bioinformatics § Gene. Weaver § my. Grid § Wednesday • Norms • Pitfalls

Agent Technology: Enabling Next Generation Computing A Roadmap for Agent Based Computing Michael Luck, University of Southampton, UK mml@ecs. soton. ac. uk

Overview § § § § What are agents? Agent. Link and the Roadmap Current state-of-the-art Short, medium and long-term predictions Technical challenges Community challenges Application Opportunities

What is an agent? § A computer system capable of flexible, autonomous (problem-solving) action, situated in dynamic, open, unpredictable and typically multi-agent domains.

What is an agent? § A computer system capable of flexible, autonomous (problem-solving) action, situated in dynamic, open, unpredictable and typically multi-agent domains. § control over internal state and over own behaviour

What is an agent? § A computer system capable of flexible, autonomous (problem-solving) action, situated in dynamic, open, unpredictable and typically multi-agent domains. § experiences environment through sensors and acts through effectors

What is an agent? § A computer system capable of flexible, autonomous (problem-solving) action, situated in dynamic, open, unpredictable and typically multi-agent domains. § reactive: respond in timely fashion to environmental change § proactive: act in anticipation of future goals

Multiple Agents In most cases, single agent is insufficient • no such thing as a single agent system (!? ) • multiple agents are the norm, to represent: § natural decentralisation § multiple loci of control § multiple perspectives § competing interests

Agent Interactions § Interaction between agents is inevitable • to achieve individual objectives, to manage interdependencies § Conceptualised as taking place at knowledgelevel • which goals, at what time, by whom, what for § Flexible run-time initiation and response • cf. design-time, hard-wired nature of extant approaches

Agent. Link and the Roadmap

What is Agent. Link? § Open network for agent-based computing. § Agent. Link II started in August 2000. § Intended to give European industry a head start in a crucial new area of IT. § Builds on existing activities from Agent. Link (1998 -2000)

Agent. Link Goals § Competitive advantage through promotion of agent systems technology § Improvement in standard, profile, industrial relevance of research in agents § Promote excellence of teaching and training § High quality forum for R&D

What does Agent. Link do? § Industry action • gaining advantage for Euro industry § Research coordination • excellence & relevance of Euro research § Education & training • fostering agent skills § Special Interest Groups • focused interactions § Information infrastructrure • facilitating Agent. Link work

The Roadmap: Aims § A key deliverable of Agent. Link II § Derives from work of Agent. Link SIGs § Draws on Industry and Research workpackages § Aimed at policy-makers, funding agencies, academics, industrialists § Aims to focus future R&D efforts

Special Interest Groups § Agent-Mediated Electronic Commerce § Agent-Based Social Simulation § Methodologies and Software Engineering for Agent Systems § Intelligent Information Agents § Intelligent and Mobile Agents for Telecoms and the Internet § Agents that Learn, Adapt and Discover § Logic and Agents

The Roadmap: Process § Core roadmapping team: • Michael Luck • Peter Mc. Burney • Chris Preist § § Inputs from SIGs: area roadmaps Specific reviews Wide consultation exercise Collation and integration

State of the art

Views of Agents To support next generation computing through facilitating agent technologies § As a metaphor for the design of complex, distributed computational systems § As a source of technologies § As simulation models of complex realworld systems, such as in biology and economics

Agents as Design § § § Agent oriented software engineering Agent architectures Mobile agents Agent infrastructure Electronic institutions

Agent technologies § § § § Multi-agent planning Agent communication languages Coordination mechanisms Matchmaking architectures Information agents and basic ontologies Auction mechanism design Negotiation strategies Learning

Links to other disciplines § § § Philosophy Logic Economics Social sciences Biology

Application and Deployment § Assistant agents § Multi-agent decision systems § Multi-agent simulation systems § IBM, HP Labs, Siemens, Motorola, BT § Lost Wax, Agent Oriented Software, Whitestein, Living Systems, i. SOCO

The Roadmap Timeline

Dimensions § § § Sharing of knowledge and goals Design by same or diverse teams Languages and interaction protocols Scale of agents, users, complexity Design methodologies

Current situation § One design team § Agents sharing common goals § Closed agent systems applied in specific environment § Ad-hoc designs § Predefined communications protocols and languages § Scalability only in simulation

Short term to 2005 § Fewer common goals § Use of semi-structured agent communication languages (such as FIPA ACL) § Top-down design methodologies such as GAIA § Scalability extended to predetermined and domain-specific environments

Medium term 2006 -2008 Design by different teams Use of agreed protocols and languages Standard, agent-specific design methodologies Open agent systems in specific domains (such as in bioinformatics and e-commerce) § More general scalability, arbitrary numbers and diversity of agents in each such domain § Bridging agents translating between domains § §

Long Term 2009§ Design by diverse teams § Truly-open and fully-scalable multi-agent systems § Across domains § Agents capable of learning appropriate communications protocols upon entry to a system § Protocols emerging and evolving through actual agent interactions.

The Roadmap Timeline

Technological Challenges

Technological Challenges § Increase quality of agent systems to industrial standard § Provide effective agreed standards to allow open systems development § Provide infrastructure for open agent communities § Develop reasoning capabilities for agents in open environments

Technological Challenges § Develop agent ability to adapt to changes in environment § Develop agent ability to understand user requirements § Ensure user confidence and trust in agents

Industrial Strength Software § Fundamental obstacle to take-up is lack of mature software methodology • Coordination, interaction, organisation, society joint goals, plans, norms, protocols, etc • Libraries of … § agent and organisation models § communication languages and patterns § ontology patterns § CASE tools § AUML is one example

Industrial Strength Software

Agreed Standards § FIPA and OMG • Agent platform architectures • Semantic communication and content languages for messages and protocols • Interoperability • Ontology modelling § Public libraries in other areas will be required

Agreed Standards

Semantic Infrastructure for Open Communities § Need to understand relation of agents, databases and information systems § Real world implications of information agents § Benchmarks for performance § Use new web standards for structural and semantic description § Services that make use of such semantic representations

Semantic Infrastructure for Open Communities § Ontologies • DAML+OIL • UML • OWL § § Timely covergence of technologies Generic tool and service support Shared ontologies Semantic Web community exploring many questions

Semantic Infrastructure for Open Communities

Reasoning in Open Environments § Cannot handle issues inherent in open multi-agent systems • • Heterogeneity Trust and accountability Failure handling and recovery Societal change § Domain-specific models of reasoning

Reasoning in Open Environments § Coalition formation § Dynamic establishment of virtual organisations § Demanded by emerging computational infrastructure such as • Grid • Web Services • e. Business workflow systems

Reasoning in Open Environments § Negotiation and argumentation • Some existing work but currently in infancy § Need to address • • • Rigorous testing in realistic environments Overarching theory or methodology Efficient argumentation engines Techniques for user preference specification Techniques for user creation and dissolution of virtual organisations

Reasoning in Open Environments

Learning Technologies § Ability to understand user requirements • Integration of machine learning • XML profiles § Ability to adapt to changes in environment • Multi-agent learning is far behind single agent learning • Personal information management raises issues of privacy § Relationship to Semantic Web

Learning Technologies

Trust and Reputation § User confidence § Trust of users in agents • Issues of autonomy • Formal methods and verification § Trust of agents in agents • Norms • Reputation • Contracts

Trust and Reputation

Challenges for the Agent Community

Community Organisation § Leverage underpinning work on similar problems in Computer Science: Object technology, software engineering, distributed systems § Link with related areas in Computer Science dealing with different problems: Artificial life, uncertainty in AI, mathematical modelling

Community Organisation § Extend and deepen links with other disciplines: Economics, logic, philosophy, sociology, etc § Encourage industry take-up: Prototypes, early adopters, case-studies, best practice, early training

Existing software technology § Build bridges with distributed systems, software engineering and object technology. § Develop agent tools and technologies on existing standards. § Engage in related (lower level) standardisation activities (UDDI, WSDL, WSFL, XLANG, OMG CORBA). § Clarify relationships between agent theories and abstract theories of distributed computation.

Different problems from related areas § Build bridges to artificial life, robotics, Uncertainty in AI, logic programming and traditional mathematical modelling. § Develop agent-based systems using hybrid approaches. § Develop metrics to assess relative strengths and weakness of different approaches.

Prior results from other disciplines § Maintain and deepen links with economics, game theory, logic, philosophy and biology. § Build new connections with sociology, anthropology, organisation design, political science, marketing theory and decision theory.

Encourage agent deployment § Build prototypes spanning organisational boundaries (potentially conflicting). § Encourage early adopters of agent technology, especially ones with some risk. § Develop catalogue of early adopter case studies, both successful and unsuccessful. § Provide analyses of reasons for success and failure cases.

Encourage agent deployment § Identify best practice for agent oriented development and deployment. § Support standardisation efforts. § Support early industry training efforts. § Provide migration paths to allow smooth evolution of agent-based solutions, from today’s solutions,

Application Opportunities

Application Opportunities § Ambient Intelligence § Bioinformatics and Computational Biology § Grid Computing § Electronic Business § Simulation § Semantic Web

Ambient Intelligence § Pillar of European Commission’s IST vision § Also developed by Philips in long-term vision § Three parts • Ubiquitous computing • Ubiquitous communication • Intelligent user interfaces § Thousands on mobile and embedded devices interacting to support user-centred goals and activity

Ambient Intelligence § Suggests a component-oriented world populated by agents • • Autonomy Distribution Adaptation Responsiveness § Demands • Virtual organisations • Infrastructure • Scalability

Bioinformatics § Information explosion in genomics and proteomics § Distributed resources include databases and analysis tools § Demands automated information gathering and inference tools § Open, dynamic and heterogeneous § Examples: Geneweaver, my. Grid

Grid Computing § Support for large scale scientific endeavour § More general applications with large scale information handling, knowledge management, service provision § Suggests virtual organisations and agents § Future model for service-oriented environments

Electronic Business § Agents currently used in first stage – merchant discovery and brokering § Next step is real trading – negotiating deals and making purchases § Potential impact on the supply chain § Rise in agent-mediated auctions expected • Agents recommend • But agents do not yet authorise agreements

Electronic Business § Short term: travel agents, etc • TAC is a driver § Long term: full supply chain integration § At start of 2001, there were • 1000 public e. Markets • 30, 000 private exchange

Simulation § Education and training § Scenario exploration § Entertainment

The Two Towers § Thousands of agents simulated using the MASSIVE system § Realistic behaviour for battle scenes § Initial versions included characters running away! § Previous use of computational characters did not use agent behaviour (eg Titanic).

Current State § Pivotal role in contributing to broader visions of Ambient Intelligence, Grid Computing, Semantic Web, etc. § European strength is broad and deep § Still requires integration, needs to avoid fragmentation, needs effective coordination § Needs to support industry take-up and innovation

For more information. . . Dr Michael Luck Department of Electronics and Computer Science University of Southampton SO 17 1 BJ United Kingdom Feedback sought: please send feedback! Roadmap: www. agentlink. org/roadmap

The Book

The CD

The Agent Portal www. agentlink. org