Designing participatory systems prof.dr. F.M.T. Brazier - - PowerPoint PPT Presentation

Designing participatory systems

prof.dr. F.M.T. Brazier Participatory Systems Initiative, Systems Engineering

Networks in which coordination is leading

Systems for which global organisation is based on local coördination

Coordination based on local situation

• own rules of engagement and agreement;
• own requirements for trust and integrity;
• own empowerment - autonomy, self-organisation

Networked systems

Networked systems

Designed for

• engagement and agreement, communication;
• trust and integrity;
• empowerment - autonomy, self-organisation

Systems for which different levels of self-

• rganisation and interaction are required

within and between different types of networks and institutions

Participatory Systems – 3 levels of structures and networks

• Social: refers to social, economical, political and

cultural dynamics

• Distributed ICT: refers to technologies that enable

large scale distributed self-organizing processes, information exchange, aggregation and clustering.

• Infrastructure: refers to the physical components/

networks of a system.

Social

Distributed ICT Physical

Social layer social structures, human-network interaction, governance, institutions Distributed ICT

• verlays, clustering,

virtual organisations Physical networks Physical infrastructure

For example energy systems – large scale distributed …. Case for adaptive autonomous systems …… agents

Social technological self organisation at the level

• f the prosumer, based on communication

Traditional energy market Energy flows one direction from a few big producers to many small consumers (and industries).

• The situation is changing
• Distributed (green) generation

wind turbines, solar panels, etc.

• Consumer becomes producer.

• Role power companies is

changing;

• Ability for consumers to sell

power to each other.

• Each entity is autonomous and

can be represented by a software agent;

• Agents can operate semi-

autonomously to negotiate agreements to buy and sell energy.

• But the real market is more

complicated;

• More agents, more

communication.

• But the real market is more

complicated;

• More agents, more

communication.

• Smaller dynamic organisations

are more effective.

• Organisations acquire

autonomy.

• Organisations negotiate with

each other.

• Organisations negotiate with

each other and change

Energy system has become a participative social-technical system !

Participatory layer

• user preferences;
• trust;
• reputation;
• policy legislation;
• self management.

}

Agent layer

• communication;
• negotiation;
• clustering;
• self management;

}

Physical layer

• houses, wind turbines;
• power lines;
• solar panels;

}

Participation

Design of participatory systems mandates

Design for trust and integrity: transparency, privacy, integrity, security, identifiability, traceability, accessibility, proportionality

• Design for empowerment:
• Design for engagement:

Requires design based on

Design for trust and integrity: transparency, privacy, integrity, security, identifiability, traceability, accessibility, proportionality Design for empowerment: autonomy, self-management and self-regulation, emergence Design for engagement:

Requires design based on

Design for trust and integrity: transparency, privacy, integrity, security, identifiability, traceability, accessibility, proportionality Design for empowerment: autonomy, self-management and self-regulation, emergence Design for engagement: presence, enactment, communication, awareness, co-creation

Taking responsibility, participating… …requires trust, awareness and the ability to act.

Social networks Distributed ICT networks Physical networks

Within 3 layers

Essential to critical infrastructures

Social networks – social structures, institutions, ..

horizontal governance polycentricity new distributed markets regulatory frameworks Situational awareness

presence design

Merging realities

The focus of Social Networks and coordination

Understanding human participation in social technical systems in merging realities: presence, social structures, governance, communities, self-organisation, risks, incentives, ….

Design principles (Ostrom) for local natural resource management

• Clearly defined boundaries
• Rules appropriate for provisioning in local context
• Collective choice decision making
• Effective monitoring
• Graduated sanctions
• Mechanisms for conflict resolutions
• Self-determination of community
• Multiple layers of nested enterprises

Systems Engineering,

To trust…

Witness presence: Caroline Nevejan Graph: Chin-Lien Chen

Teletrust: Design for trust

TeleTrust: Design for human-system interaction in new types of socio-techno eco systems

A few of the challenges

• monitoring vs privacy, ownership

intrusion detection, privacy preserving measures,

• communities
• orchestration of emergent behaviour
• formal and informal regulations - national/EU/W certification,

quality

• trust mechanisms in merging realities

http://www.flickr.com/photos/mshandro/35000426/lightbox/

Distributed ICT networks and coordination

Distributed ICT networks – self-healing, overlays, ..

dynamic clustering Dynamic reconfiguration Load balancing SLAs, distributed monitoring Virtual power stations Distributed accounting dynamic aggregation

The focus of participation at the level

• f Distributed ICT

distributed social overlays/communities, load balancing distributed aggregation, distributed SLAs, distributed monitoring, distributed clustering Robustness, Resilience, Ability to Adapt, Security, Cascading effects

autonomous adaptive systems

autonomic computing systems p2p systems, embedded systems ……. multi-agent systems in disguise

In virtual organisations of autonomous systems

Physically distributed groups of heterogeneous autonomous systems possibly with different levels of accessibility, authorisation, authentication that interact with each other to collectively or individually accomplish one or more tasks

Through interaction dynamic groups of systems emerge ..

resource/load balancing energy management, data centre management, crisis management swarm applications supply chain management traffic management

Four challenges within distributed ICT

• 1. multi-level communication structures
• 2. multi-level control and aggregation
• 3. multi-level commitment

Key challenge 1: communication structures

Structure of a communication network

• dynamic overlay structures within each level
• dynamic interaction between levels

Eg How structured does a network need to be?

Hierarachical network structure

Semi-structured communication network

Clustered network structure

unstructured network structures

layered network structures

Example: dynamic robust tree overlays

Pournaras, 2009

Key challenge 2: Control structures and aggregation

Regulating division of responsibility:

Virtual Power Stations: which level of aggregation makes most sense?

VP

Power grid

VP VP VP

Power grid

Virtual Power Stations: which level of aggregation makes most sense?

Ogston 2009

Challenge the future

Delft University of Technology

Distributed Coordination

Water heater Freezer Fridge Air-conditioner Intelligent Agents Tree overlay for aggregation

Evangelos Pournaras, March 2009

Key challenge 3: commitments

Protocols, SLAs, WSAS, Well-structured, well-defined, OGF? State?

For example – Web Service Agreement Specs

C C P P P C

template agreement request agreement

WSAS based negotiation

A P P A A P My requirements are … This is the contract I’m offering you A P Accept contract A p What do you have to offer These services are currently available

WSAS based mediated negotiation

Implemented for energy domain, computer resources

power plant Wind turbine solar panels

mobach, 2005

All very promising developments

For large-scale, autonomous, interactive, self-managing systems Needs extending for multi-round negotiation, multi- attribute negotiation, multi-party negotiation

challenges

Integrity - (Local) aggregation and dissemination of information Accountability - distributed SLA negotiation, distributed monitoring, identity mgt Containment – resilience - avoiding cascading effects Reliabilty – self organisation, self management Scalability – millions of devices

At all 3 levels

Physical networks - power lines, resources, ….

Electric vehicles DC micro grids Self-healing distribution networks Physical chain Topology optimization Intelligent control

greenhouses production transportation

street

Focus of physical networks and coordination

Health of the physical network, design of robust topologies, resources and storage, … Supply chain operation

challenges

Scalability Robustness Resilience Maintainability QoS, timeliness ……

Essential for critical infrastructures – US Airforce

“… ¡must ¡move ¡even ¡further ¡toward ¡autonomous ¡systems ¡linked ¡to ¡each ¡

• ther ¡and ¡to ¡service ¡members ¡through ¡cyberspace ¡to ¡deliver ¡increased ¡

capability ¡with ¡decreased ¡cost” ¡ ¡ “effective ¡merger ¡of ¡man ¡and ¡machine ¡capabilities: ¡augmented ¡cognition, ¡ integrated ¡cyber ¡human ¡systems” ¡ ¡ “capability ¡to ¡“island” ¡installations ¡while ¡maintaining ¡operational ¡surety ¡ and ¡security.”

Impression of recent work

Social networks and coordination

CART – Civilian Alert Real Time (National Police, the Hague City, TNO, CGI) SamenMarkt: Restoring trust in the horticultural fresh food market using multi-agent system technology (Wageningen, growers, traders, ….) NGI New Governance Models for Next Generation Infrastructures (Alliander, Thales) Supply-demand chain coordination (Iran) l

Social networks and coordination

Cities as drivers of social change (CIVIS, EU) Securing the European Electricity Supply Against Malicious and accidental thrEats (SESAME, EU) BRIDGE: sustainaBle uRban plannIng Decision support accountinG for urban mEtabolism (EU FP7 IP) On the Spot (National Forensic Institute, National Police, Fire Department Rijnland)

Social networks and coordination

Can I touch you online? (Lancel & Maat) Future Scenarios and Smart Energy Systems Experience Labs (EIT ICT Labs) Slim Verbinden (Agentschap NL) Self-Managed Dynamic Institutions in crisis situations (Thales, CSI the Hague – mediated reality (NFI, National Police, …) Freight logistics

Distributed ICT networks and coordination

SES NWO: Increasing the Robustness of Smart Grids through distributed energy generation: a complex network approach (Brazier, Kooij, Warnier, Smit) NWO STASCADE: Stable and scalable decentralized power balancing systems using adaptive clustering (Brazier, la Poutre, Warnier) NWO-India ADREM : Adaptive clustering for Decentralized Resilient Energy Management (Brazier, la Poutre, Warnier)

Distributed ICT networks and coordination

European Virtual Smart Grid Lab (EIT ICT Labs) NGI Self-Managed Dynamic Institutions in Power Grids: Sharing the cost of reliability ( Tennet, Alliander, Kema, Thales) STARS Sensor Technology applied across sensor networks (NL wide) NLnet Design and Management of Networked Autonomous Systems

Design?

73

How?

In addition to literature review, interviews, enquetes, ….

Gaming

Exploration - emergent behaviour, insights, effects of design choices Situational awareness - how, what, why (Training)

75

Distributed simulation and Emulation

Model and study system behaviour during different phases of system design, what-if, … Explore different design choices during prototyping

• Fully distributed vs central vs semi-distributed
• One-one negotation vs broadcasting vs multi-casting
• Cluster size

With stakeholders – serious games

76

Design of distributed systems

Single machine simulation agents

Design of distributed systems

Single machine simulation

Distributed simulation

AgentScape middleware platform supports both simulation and emulation of large-scale heterogeneous agent systems.

Challenge the future

Delft University of Technology

What are the steps in the development cycle?

Program Time Network Application Application Setting Theoretical analysis modeled modeled modeled modeled modeled 1a Synchronized simulation real modeled modeled modeled modeled 1b Asynchronous Simulation real real modeled modeled modeled 2 Emulation real real real modeled modeled 3 Demonstrator real real real real modeled Application deployment real real real real real

October 21, 2015 79 Dynamic Adaptive Systems Group, Systems Engineering, TPM

AgentScape: designed to support 4 phases of design and deployment of large scale MAS

Agents Single machine simulation Distributed emulation Deployment

Roadmap AgentScape – TU Delft and Thales NL R&D

80

AgentScape* is a platform designed to this purpose…

Framework for

• scalable,
• interactive,
• secure,
• robust,
• interoperable
• distributed autonomous systems/agents
• supporting service agreement negotiation
• and service agreement enforcement

*AgentScape development is currently a joint endeavour with Thales NL

AgentScape

AgentScape kernel

host manager location manager agent server

AgentScape kernel interface

Web service gateway host manager agent server

AgentScape middleware services AgentScape location

AgentScape kernel

host manager location manager agent server AgentScape kernel interface web service gateway host manager agent server

AgentScape middleware services Agentscope

Network/Environment interface

October 21, 2015 83

Thank you for your attention! www.participatorysystems.org

Design: Cok Francken

Need for a new design paradigm

Service Participation Product

Participatory Systems – 3 levels of structures and networks

• Social: refers to social, economical, political and

cultural dynamics

• Distributed ICT: refers to technologies that enable

large scale distributed self-organizing processes, information exchange, aggregation and clustering.

• Infrastructure: refers to the physical components/

networks of a system.

Social

Distributed ICT Physical

MAS

• Transportation networks – logistics: Exploration of future

multi-modal scenarios with stakeholders

• Smart Grid: Local load balancing, self-managed distribution

networks based on the health of the physical network

• Safety Management: crises, local communities …
• Market dynamics: supply demand networks
• Traffic management: flow management, lights, ramps
• Smart sensors
• Advance health care management
• Radar coordination