SlideShare ist ein Scribd-Unternehmen logo
1 von 31
Downloaden Sie, um offline zu lesen
Outline Discovery Strategy Promotion Techniques Results Conclusions
Strategies for Cooperation Emergence
in Distributed Service Discovery
E. del Val M. Rebollo V. Botti
Univ. Politècnica de València (Spain)
COREDEMA ’13
Salamanca, May 2013
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Promoting Cooperation
Motivation
There are scenarios in decentralized systems in which cooperation
plays a central role
agents connected in networks
bounded rationality
heterogeneous, self-interested agents
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Our Proposal
The challenge
Obtain an emergent, cooperative global behavior even when
cooperators are a minority, from local decisions.
What is done. . .
a network structure that ensures navigation and efficiency
structural changes to isolate undesired agents
variable incentives to promote cooperation
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Outline
1 Outline
2 Discovery Strategy
3 Isolated Cooperation Promotion Techniques
4 Combined Cooperation Model
5 Results
6 Conclusions
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Agent Network Model
A = {1, ..., n} a set of agents connected in a
undirected network G, where N(i) denotes the neighbors of
agent i
each agent plays a role ri and offers a service si
agents have an initial behavior: cooperative (c) or not
cooperative (nc)
each agent has an initial budget b
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Service Discovery
Purpose
Locate in the network a similar enough service offer by a concrete
role
qt
i = {stg , rtg , TTL, ε, {}}
stg required semantic service description
rtg organizational role the target agent should play
TTL: time to live
ε similarity threshold in [0, 1]
{} participant list (initially empty)
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Discovery Process
i
Ri = {r1}
Si = {s1}
k
CH(k, t) = 0.5
j
CH(j, t) = 0.5
n
CH(n, t) = 0.15
A S R |N|
k Sk Rk = {r1} 5
n Sn Rn = {r2} 5
j Sj Rj = {r1} 4
v
Rt = {r5}
St = {s6}
m
Rm = {r7}
Sm = {s7}
each agent knows its
direct neighbors
query qt
i is redirected to
the most promising
neighbor
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Similarity Measure
FNi (tg) = argmaxj∈Ni
P( j, tg )
For each neighbor j, P( j, tg ) determines the probability that the
neighbor j redirects the search to the nearest network community
where there are more probabilities of finding the agent tg.
P( j, tg ) = 1 −





1 −





CH(j, tg)
k∈Ni
CH(k, tg)










kj
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Social Plasticity
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 10 20 30 40 50 60
Probabilitytomaintainthelink
Number of queries that were forwarded to other links
n = 2
n = 4
n = 6
rewiring action λ to avoid
non-cooperative agents
decay function using a
sigmoid
d parameter establishes
benevolence of the agent
Pdecay (rqij) = 1
1+e
−(rqij −d)
n
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Social Plasticity Effects
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Social Plasticity Effects
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Incentives Effect
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Social Plasticity and Incentives
When a neighbor j receives a query qt
i , it has a set of possible
actions Ac = {ρ, ∞, 1, 2, ..., ki , ∅, λ}, where:
ρ is asking for a service
∞ is providing the service
{1, ..., ki } is forwarding the query to one of its neighbors ∈ Ni
∅ is doing nothing
λ rewiring a link
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Action Selection
Action Condition
at
i = ∞ if |CH(i, tg)| ≥ ε
at
i = ∅ if |CH(i, tg)| < ε ∧
at−1
j = ∅, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = j if |CH(i, tg)| < ε ∧
at−1
j = 0, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = λ if at−1
i = j ∧
at
j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop.
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Action Selection
Action Condition
at
i = ∞
if |CH(i, tg)| ≥ ε
at
i = ∅ if |CH(i, tg)| < ε ∧
at−1
j = ∅, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = j if |CH(i, tg)| < ε ∧
at−1
j = 0, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = λ if at−1
i = j ∧
at
j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop.
Do the task if agent knows how to do it
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Action Selection
Action Condition
at
i = ∞ if |CH(i, tg)| ≥ ε
at
i = ∅
if |CH(i, tg)| < ε ∧
at−1
j = ∅, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = j if |CH(i, tg)| < ε ∧
at−1
j = 0, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = λ if at−1
i = j ∧
at
j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop.
Do nothing if the agent guess that the most promising neighbor
will no cooperate
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Action Selection
Action Condition
at
i = ∞ if |CH(i, tg)| ≥ ε
at
i = ∅ if |CH(i, tg)| < ε ∧
at−1
j = ∅, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = j
if |CH(i, tg)| < ε ∧
at−1
j = 0, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = λ if at−1
i = j ∧
at
j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop.
Forward the query if the agent guess that the most promising
neighbor will cooperate
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Action Selection
Action Condition
at
i = ∞ if |CH(i, tg)| ≥ ε
at
i = ∅ if |CH(i, tg)| < ε ∧
at−1
j = ∅, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = j if |CH(i, tg)| < ε ∧
at−1
j = 0, j ∈ argmax(CHt−1
1 , ..., CHt−1
ki
)
at
i = λ
if at−1
i = j ∧
at
j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop.
Rewire some links with a probability Pdecay if the agent is
surrounded by non-coop agents
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Costs of the Actions
ut
i (at
i ) =



−β if at
i = ρ
p if at
i = ∞
−c if at
i ∈ {1, 2, ..., ki }
0 if at
i = ∅ ∧ t ≤ t : at
i ∈ {1, 2, ...ki }
α if at
i = ∅ ∧ ∃t ≤ t : at
i ∈ {1, 2, ..., ki } ∧ ∃j ∈ A : at
j = ∞
−γ if at
i = λ
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Incentives Policy
uniformly distributed
System the system provides incentives
Fixed the agent that request the service pays for it
base on a criterion
Path depends on the length of the path
SimDg the more similar the higher reward
InvSimDg the less similar the higher reward
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Experimental Parameters
network size: 1 000 agents
average degree of connection: 2.5
similarity threshold ε = 0.75
TTL = 100
initial budget: 100
40 % cooperative - 60 % non cooperative
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Budget Distribution
Incentives
0
200
400
600
800
1000
1200
1400
1600
2 4 6 8 10 12 14 16 18 20
budget
degree of connection
Fixed Path Sim InvSim
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Budget Distribution
Incentives
0
200
400
600
800
1000
1200
1400
1600
2 4 6 8 10 12 14 16 18 20
budget
degree of connection
Fixed Path Sim InvSim
Incentives + Social Plasticity
0
200
400
600
800
1000
1200
1400
1600
2 4 6 8 10 12 14 16 18 20
budget degree of connection
Fixed Path Sim InvSim
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Cooperative Behavior Rate
Incentives
0
200
400
600
800
1000
2 4 6 8 10 12 14 16 18
coop
snapshot
Fixed
Path
Sim
InvSim
System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Cooperative Behavior Rate
Incentives
0
200
400
600
800
1000
2 4 6 8 10 12 14 16 18
coop
snapshot
Fixed
Path
Sim
InvSim
System
Incentives + Social Plasticity
0
200
400
600
800
1000
2 4 6 8 10 12 14 16 18
coop snapshot
Fixed
Path
Sim
InvSim
System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Success Rate
Incentives
0
20
40
60
80
100
2 4 6 8 10 12 14 16 18
%successfulsearches
snapshot
Fixed Path Sim InvSim System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Success Rate
Incentives
0
20
40
60
80
100
2 4 6 8 10 12 14 16 18
%successfulsearches
snapshot
Fixed Path Sim InvSim System
Incentives + Social Plasticity
0
20
40
60
80
100
2 4 6 8 10 12 14 16 18
%successfulsearches snapshot
Fixed
Path
Sim
InvSim
System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Path Length
Incentives
0
20
40
60
80
100
2 4 6 8 10 12 14 16 18
steps
snapshot
Fixed Path Sim InvSim System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Path Length
Incentives
0
20
40
60
80
100
2 4 6 8 10 12 14 16 18
steps
snapshot
Fixed Path Sim InvSim System
Incentives + Social Plasticity
0
20
40
60
80
100
2 4 6 8 10 12 14 16 18
steps snapshot
Fixed
Path
Sim
InvSim
System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Num. of Broken Links (Rewired)
0
200
400
600
800
1000
2 4 6 8 10 12 14 16 18
budget
snapshot
Fixed Path Sim InvSim System
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery
Outline Discovery Strategy Promotion Techniques Results Conclusions
Conclusions
What we’ve done
To combine structural changes (social plasticity) with different
incentives policies in a decentralized service discovery scenario with
local search.
What we’ve got
variable incentives work better than homogenous ones
combination of mechanisms promotes cooperation in scenarios
in which |nc| > |c|
it increases the performance of the agents
reduces the average path length
increases the success rate
M. Rebollo et al. (UPV) COREDEMA’13
Strategies for Cooperation Emergence in Distributed Service Discovery

Weitere ähnliche Inhalte

Andere mochten auch

Andere mochten auch (17)

Análisis dinámico de redes sociales en diferentes eventos
Análisis dinámico de redes sociales en diferentes eventosAnálisis dinámico de redes sociales en diferentes eventos
Análisis dinámico de redes sociales en diferentes eventos
 
Análisis de redes comercio mediante procesos de consenso
Análisis de redes comercio mediante procesos de consensoAnálisis de redes comercio mediante procesos de consenso
Análisis de redes comercio mediante procesos de consenso
 
Generalized consensus process in dynamic networks
Generalized consensus process in dynamic networksGeneralized consensus process in dynamic networks
Generalized consensus process in dynamic networks
 
Procesos de consenso en redes dinámicas
Procesos de consenso en redes dinámicasProcesos de consenso en redes dinámicas
Procesos de consenso en redes dinámicas
 
El poder de las redes
El poder de las redesEl poder de las redes
El poder de las redes
 
Presentacion AIWS
Presentacion AIWSPresentacion AIWS
Presentacion AIWS
 
Cómo crear mi PLE
Cómo crear mi PLECómo crear mi PLE
Cómo crear mi PLE
 
Comunica 2016
Comunica 2016Comunica 2016
Comunica 2016
 
Guía para el uso de redes sociales en el aprendizaje inverso
Guía para el uso de redes sociales en el aprendizaje inversoGuía para el uso de redes sociales en el aprendizaje inverso
Guía para el uso de redes sociales en el aprendizaje inverso
 
Analysis of the Evolution of Events on Online Social Networks
Analysis of the Evolution of Events on Online Social NetworksAnalysis of the Evolution of Events on Online Social Networks
Analysis of the Evolution of Events on Online Social Networks
 
Supportive consensus for smart grid management
Supportive consensus for smart grid managementSupportive consensus for smart grid management
Supportive consensus for smart grid management
 
La (poca) soledad del corredor de fondo
La (poca) soledad del corredor de fondoLa (poca) soledad del corredor de fondo
La (poca) soledad del corredor de fondo
 
Consensus in Smart Grids for Decentralized Energy Management
Consensus in Smart Grids for Decentralized Energy ManagementConsensus in Smart Grids for Decentralized Energy Management
Consensus in Smart Grids for Decentralized Energy Management
 
Decentralized Group AHP in Multilayer Networks by Consensus
Decentralized Group AHP in Multilayer Networks by ConsensusDecentralized Group AHP in Multilayer Networks by Consensus
Decentralized Group AHP in Multilayer Networks by Consensus
 
Análisis de sentimientos en Twitter mediante HMM
Análisis de sentimientos en Twitter mediante HMMAnálisis de sentimientos en Twitter mediante HMM
Análisis de sentimientos en Twitter mediante HMM
 
U-Tool: A Urban-Toolkit for enhancing city maps through citizens’ activity
U-Tool: A Urban-Toolkit for enhancing city maps through citizens’ activityU-Tool: A Urban-Toolkit for enhancing city maps through citizens’ activity
U-Tool: A Urban-Toolkit for enhancing city maps through citizens’ activity
 
Meilleur travail du monde
Meilleur travail du mondeMeilleur travail du monde
Meilleur travail du monde
 

Ähnlich wie Strategies for Cooperation Emergence in Distributed Service Discovery

On the Effectiveness of Offline RL for Dialogue Response Generation.pdf
On the Effectiveness of Offline RL for Dialogue Response Generation.pdfOn the Effectiveness of Offline RL for Dialogue Response Generation.pdf
On the Effectiveness of Offline RL for Dialogue Response Generation.pdf
Po-Chuan Chen
 
Mechanisms to Promote Cooperation in Decentralized Services
Mechanisms to Promote Cooperation in Decentralized ServicesMechanisms to Promote Cooperation in Decentralized Services
Mechanisms to Promote Cooperation in Decentralized Services
Miguel Rebollo
 
Lec14: Evaluation Framework for Medical Image Segmentation
Lec14: Evaluation Framework for Medical Image SegmentationLec14: Evaluation Framework for Medical Image Segmentation
Lec14: Evaluation Framework for Medical Image Segmentation
Ulaş Bağcı
 
A software fault localization technique based on program mutations
A software fault localization technique based on program mutationsA software fault localization technique based on program mutations
A software fault localization technique based on program mutations
Tao He
 
Updated paper in latex(1)
Updated paper in latex(1)Updated paper in latex(1)
Updated paper in latex(1)
Richa Shukla
 

Ähnlich wie Strategies for Cooperation Emergence in Distributed Service Discovery (20)

Deep Reinforcement Learning Through Policy Optimization, John Schulman, OpenAI
Deep Reinforcement Learning Through Policy Optimization, John Schulman, OpenAIDeep Reinforcement Learning Through Policy Optimization, John Schulman, OpenAI
Deep Reinforcement Learning Through Policy Optimization, John Schulman, OpenAI
 
On the Effectiveness of Offline RL for Dialogue Response Generation.pdf
On the Effectiveness of Offline RL for Dialogue Response Generation.pdfOn the Effectiveness of Offline RL for Dialogue Response Generation.pdf
On the Effectiveness of Offline RL for Dialogue Response Generation.pdf
 
Link Discovery Tutorial Part I: Efficiency
Link Discovery Tutorial Part I: EfficiencyLink Discovery Tutorial Part I: Efficiency
Link Discovery Tutorial Part I: Efficiency
 
Presentation of the unbalanced R package
Presentation of the unbalanced R packagePresentation of the unbalanced R package
Presentation of the unbalanced R package
 
Matrix Factorization Technique for Recommender Systems
Matrix Factorization Technique for Recommender SystemsMatrix Factorization Technique for Recommender Systems
Matrix Factorization Technique for Recommender Systems
 
A Study of Smoothing Methods for Relevance-Based Language Modelling of Recomm...
A Study of Smoothing Methods for Relevance-Based Language Modelling of Recomm...A Study of Smoothing Methods for Relevance-Based Language Modelling of Recomm...
A Study of Smoothing Methods for Relevance-Based Language Modelling of Recomm...
 
Adaptive Machine Learning for Credit Card Fraud Detection
Adaptive Machine Learning for Credit Card Fraud DetectionAdaptive Machine Learning for Credit Card Fraud Detection
Adaptive Machine Learning for Credit Card Fraud Detection
 
CoopLoc Technical Presentation
CoopLoc Technical PresentationCoopLoc Technical Presentation
CoopLoc Technical Presentation
 
Mechanisms to Promote Cooperation in Decentralized Services
Mechanisms to Promote Cooperation in Decentralized ServicesMechanisms to Promote Cooperation in Decentralized Services
Mechanisms to Promote Cooperation in Decentralized Services
 
Lec14: Evaluation Framework for Medical Image Segmentation
Lec14: Evaluation Framework for Medical Image SegmentationLec14: Evaluation Framework for Medical Image Segmentation
Lec14: Evaluation Framework for Medical Image Segmentation
 
Racing for unbalanced methods selection
Racing for unbalanced methods selectionRacing for unbalanced methods selection
Racing for unbalanced methods selection
 
Efficiency of Power Distribution Companies in Pakistan (Application of Non Pa...
Efficiency of Power Distribution Companies in Pakistan (Application of Non Pa...Efficiency of Power Distribution Companies in Pakistan (Application of Non Pa...
Efficiency of Power Distribution Companies in Pakistan (Application of Non Pa...
 
A software fault localization technique based on program mutations
A software fault localization technique based on program mutationsA software fault localization technique based on program mutations
A software fault localization technique based on program mutations
 
Supervised-Learning Link Recommendation in the DBLP co-authoring network
Supervised-Learning Link Recommendation in the DBLP co-authoring networkSupervised-Learning Link Recommendation in the DBLP co-authoring network
Supervised-Learning Link Recommendation in the DBLP co-authoring network
 
DEA
DEADEA
DEA
 
50120130406014
5012013040601450120130406014
50120130406014
 
Updated paper in latex(1)
Updated paper in latex(1)Updated paper in latex(1)
Updated paper in latex(1)
 
Towards Evidence Terminological Decision Tree
Towards Evidence Terminological Decision TreeTowards Evidence Terminological Decision Tree
Towards Evidence Terminological Decision Tree
 
Newtonian Law Inspired Optimization Techniques Based on Gravitational Search ...
Newtonian Law Inspired Optimization Techniques Based on Gravitational Search ...Newtonian Law Inspired Optimization Techniques Based on Gravitational Search ...
Newtonian Law Inspired Optimization Techniques Based on Gravitational Search ...
 
2015-03-31_MotifGP
2015-03-31_MotifGP2015-03-31_MotifGP
2015-03-31_MotifGP
 

Mehr von Miguel Rebollo

Mehr von Miguel Rebollo (19)

GTG-CoL: A Decentralized Federated Learning Based on Consensus for Dynamic N...
 GTG-CoL: A Decentralized Federated Learning Based on Consensus for Dynamic N... GTG-CoL: A Decentralized Federated Learning Based on Consensus for Dynamic N...
GTG-CoL: A Decentralized Federated Learning Based on Consensus for Dynamic N...
 
Co-Learning: Consensus-based Learning for Multi-Agent Systems
 Co-Learning: Consensus-based Learning for Multi-Agent Systems Co-Learning: Consensus-based Learning for Multi-Agent Systems
Co-Learning: Consensus-based Learning for Multi-Agent Systems
 
Análisis de la red de autores de ciencia ficción de Clarkesworld
Análisis de la red de autores de ciencia ficción de ClarkesworldAnálisis de la red de autores de ciencia ficción de Clarkesworld
Análisis de la red de autores de ciencia ficción de Clarkesworld
 
Y sin embargo... se mueve. Dinámica de las redes complejas
Y sin embargo... se mueve. Dinámica de las redes complejasY sin embargo... se mueve. Dinámica de las redes complejas
Y sin embargo... se mueve. Dinámica de las redes complejas
 
Exámenes en grupo y pruebas de corrección como alternativas a la evaluación
Exámenes en grupo y pruebas de corrección como alternativas a la evaluaciónExámenes en grupo y pruebas de corrección como alternativas a la evaluación
Exámenes en grupo y pruebas de corrección como alternativas a la evaluación
 
Gamification. Key Concepts
Gamification. Key ConceptsGamification. Key Concepts
Gamification. Key Concepts
 
Using Distributed Risk Maps by Consensus as a Complement to Contact Tracing Apps
Using Distributed Risk Maps by Consensus as a Complement to Contact Tracing AppsUsing Distributed Risk Maps by Consensus as a Complement to Contact Tracing Apps
Using Distributed Risk Maps by Consensus as a Complement to Contact Tracing Apps
 
Distributed Ledger and Robust Consensus for Agreements
Distributed Ledger and Robust Consensus for AgreementsDistributed Ledger and Robust Consensus for Agreements
Distributed Ledger and Robust Consensus for Agreements
 
Detección de nodos tramposos en procesos de consenso en redes
Detección de nodos tramposos en procesos de consenso en redesDetección de nodos tramposos en procesos de consenso en redes
Detección de nodos tramposos en procesos de consenso en redes
 
La hora del código: ApS para fomentar el pensamiento computacional
La hora del código: ApS para fomentar el pensamiento computacionalLa hora del código: ApS para fomentar el pensamiento computacional
La hora del código: ApS para fomentar el pensamiento computacional
 
Procesos de enseñanza-aprendizaje en red
Procesos de enseñanza-aprendizaje en redProcesos de enseñanza-aprendizaje en red
Procesos de enseñanza-aprendizaje en red
 
desarrollo de competencias a través de narrativas transmedia
desarrollo de competencias a través de narrativas transmediadesarrollo de competencias a través de narrativas transmedia
desarrollo de competencias a través de narrativas transmedia
 
Distributed Group Analytical Hierarchical Process by Consensus
 Distributed Group Analytical Hierarchical Process by Consensus Distributed Group Analytical Hierarchical Process by Consensus
Distributed Group Analytical Hierarchical Process by Consensus
 
Análisis de ciudades a través de su actividad en redes sociales
Análisis de ciudades a través de su actividad en redes socialesAnálisis de ciudades a través de su actividad en redes sociales
Análisis de ciudades a través de su actividad en redes sociales
 
Análisis de datos en redes sociales
Análisis de datos en redes socialesAnálisis de datos en redes sociales
Análisis de datos en redes sociales
 
The multigent Layer for CALMeD SURF
The multigent Layer for CALMeD SURFThe multigent Layer for CALMeD SURF
The multigent Layer for CALMeD SURF
 
Narrativa transmedia en el aula
Narrativa transmedia en el aulaNarrativa transmedia en el aula
Narrativa transmedia en el aula
 
Using geo-tagged sentiment to better understand social interactions
 Using geo-tagged sentiment to better understand social interactions Using geo-tagged sentiment to better understand social interactions
Using geo-tagged sentiment to better understand social interactions
 
Transport Network Analysis for Smart Open Fleets
Transport Network Analysis for Smart Open FleetsTransport Network Analysis for Smart Open Fleets
Transport Network Analysis for Smart Open Fleets
 

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Data Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt RobisonData Cloud, More than a CDP by Matt Robison
Data Cloud, More than a CDP by Matt Robison
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
 
Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...Driving Behavioral Change for Information Management through Data-Driven Gree...
Driving Behavioral Change for Information Management through Data-Driven Gree...
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
 
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a FresherStrategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
GenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdfGenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdf
 
GenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day PresentationGenCyber Cyber Security Day Presentation
GenCyber Cyber Security Day Presentation
 
What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?What Are The Drone Anti-jamming Systems Technology?
What Are The Drone Anti-jamming Systems Technology?
 
08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men08448380779 Call Girls In Friends Colony Women Seeking Men
08448380779 Call Girls In Friends Colony Women Seeking Men
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘
 
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 

Strategies for Cooperation Emergence in Distributed Service Discovery

  • 1. Outline Discovery Strategy Promotion Techniques Results Conclusions Strategies for Cooperation Emergence in Distributed Service Discovery E. del Val M. Rebollo V. Botti Univ. Politècnica de València (Spain) COREDEMA ’13 Salamanca, May 2013 M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 2. Outline Discovery Strategy Promotion Techniques Results Conclusions Promoting Cooperation Motivation There are scenarios in decentralized systems in which cooperation plays a central role agents connected in networks bounded rationality heterogeneous, self-interested agents M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 3. Outline Discovery Strategy Promotion Techniques Results Conclusions Our Proposal The challenge Obtain an emergent, cooperative global behavior even when cooperators are a minority, from local decisions. What is done. . . a network structure that ensures navigation and efficiency structural changes to isolate undesired agents variable incentives to promote cooperation M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 4. Outline Discovery Strategy Promotion Techniques Results Conclusions Outline 1 Outline 2 Discovery Strategy 3 Isolated Cooperation Promotion Techniques 4 Combined Cooperation Model 5 Results 6 Conclusions M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 5. Outline Discovery Strategy Promotion Techniques Results Conclusions Agent Network Model A = {1, ..., n} a set of agents connected in a undirected network G, where N(i) denotes the neighbors of agent i each agent plays a role ri and offers a service si agents have an initial behavior: cooperative (c) or not cooperative (nc) each agent has an initial budget b M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 6. Outline Discovery Strategy Promotion Techniques Results Conclusions Service Discovery Purpose Locate in the network a similar enough service offer by a concrete role qt i = {stg , rtg , TTL, ε, {}} stg required semantic service description rtg organizational role the target agent should play TTL: time to live ε similarity threshold in [0, 1] {} participant list (initially empty) M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 7. Outline Discovery Strategy Promotion Techniques Results Conclusions Discovery Process i Ri = {r1} Si = {s1} k CH(k, t) = 0.5 j CH(j, t) = 0.5 n CH(n, t) = 0.15 A S R |N| k Sk Rk = {r1} 5 n Sn Rn = {r2} 5 j Sj Rj = {r1} 4 v Rt = {r5} St = {s6} m Rm = {r7} Sm = {s7} each agent knows its direct neighbors query qt i is redirected to the most promising neighbor M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 8. Outline Discovery Strategy Promotion Techniques Results Conclusions Similarity Measure FNi (tg) = argmaxj∈Ni P( j, tg ) For each neighbor j, P( j, tg ) determines the probability that the neighbor j redirects the search to the nearest network community where there are more probabilities of finding the agent tg. P( j, tg ) = 1 −      1 −      CH(j, tg) k∈Ni CH(k, tg)           kj M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 9. Outline Discovery Strategy Promotion Techniques Results Conclusions Social Plasticity 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 10 20 30 40 50 60 Probabilitytomaintainthelink Number of queries that were forwarded to other links n = 2 n = 4 n = 6 rewiring action λ to avoid non-cooperative agents decay function using a sigmoid d parameter establishes benevolence of the agent Pdecay (rqij) = 1 1+e −(rqij −d) n M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 10. Outline Discovery Strategy Promotion Techniques Results Conclusions Social Plasticity Effects M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 11. Outline Discovery Strategy Promotion Techniques Results Conclusions Social Plasticity Effects M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 12. Outline Discovery Strategy Promotion Techniques Results Conclusions Incentives Effect M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 13. Outline Discovery Strategy Promotion Techniques Results Conclusions Social Plasticity and Incentives When a neighbor j receives a query qt i , it has a set of possible actions Ac = {ρ, ∞, 1, 2, ..., ki , ∅, λ}, where: ρ is asking for a service ∞ is providing the service {1, ..., ki } is forwarding the query to one of its neighbors ∈ Ni ∅ is doing nothing λ rewiring a link M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 14. Outline Discovery Strategy Promotion Techniques Results Conclusions Action Selection Action Condition at i = ∞ if |CH(i, tg)| ≥ ε at i = ∅ if |CH(i, tg)| < ε ∧ at−1 j = ∅, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = j if |CH(i, tg)| < ε ∧ at−1 j = 0, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = λ if at−1 i = j ∧ at j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop. M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 15. Outline Discovery Strategy Promotion Techniques Results Conclusions Action Selection Action Condition at i = ∞ if |CH(i, tg)| ≥ ε at i = ∅ if |CH(i, tg)| < ε ∧ at−1 j = ∅, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = j if |CH(i, tg)| < ε ∧ at−1 j = 0, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = λ if at−1 i = j ∧ at j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop. Do the task if agent knows how to do it M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 16. Outline Discovery Strategy Promotion Techniques Results Conclusions Action Selection Action Condition at i = ∞ if |CH(i, tg)| ≥ ε at i = ∅ if |CH(i, tg)| < ε ∧ at−1 j = ∅, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = j if |CH(i, tg)| < ε ∧ at−1 j = 0, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = λ if at−1 i = j ∧ at j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop. Do nothing if the agent guess that the most promising neighbor will no cooperate M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 17. Outline Discovery Strategy Promotion Techniques Results Conclusions Action Selection Action Condition at i = ∞ if |CH(i, tg)| ≥ ε at i = ∅ if |CH(i, tg)| < ε ∧ at−1 j = ∅, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = j if |CH(i, tg)| < ε ∧ at−1 j = 0, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = λ if at−1 i = j ∧ at j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop. Forward the query if the agent guess that the most promising neighbor will cooperate M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 18. Outline Discovery Strategy Promotion Techniques Results Conclusions Action Selection Action Condition at i = ∞ if |CH(i, tg)| ≥ ε at i = ∅ if |CH(i, tg)| < ε ∧ at−1 j = ∅, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = j if |CH(i, tg)| < ε ∧ at−1 j = 0, j ∈ argmax(CHt−1 1 , ..., CHt−1 ki ) at i = λ if at−1 i = j ∧ at j = ∅ ∧ |coop| < σ, coop ⊆ Ni (g)|j is a coop. Rewire some links with a probability Pdecay if the agent is surrounded by non-coop agents M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 19. Outline Discovery Strategy Promotion Techniques Results Conclusions Costs of the Actions ut i (at i ) =    −β if at i = ρ p if at i = ∞ −c if at i ∈ {1, 2, ..., ki } 0 if at i = ∅ ∧ t ≤ t : at i ∈ {1, 2, ...ki } α if at i = ∅ ∧ ∃t ≤ t : at i ∈ {1, 2, ..., ki } ∧ ∃j ∈ A : at j = ∞ −γ if at i = λ M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 20. Outline Discovery Strategy Promotion Techniques Results Conclusions Incentives Policy uniformly distributed System the system provides incentives Fixed the agent that request the service pays for it base on a criterion Path depends on the length of the path SimDg the more similar the higher reward InvSimDg the less similar the higher reward M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 21. Outline Discovery Strategy Promotion Techniques Results Conclusions Experimental Parameters network size: 1 000 agents average degree of connection: 2.5 similarity threshold ε = 0.75 TTL = 100 initial budget: 100 40 % cooperative - 60 % non cooperative M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 22. Outline Discovery Strategy Promotion Techniques Results Conclusions Budget Distribution Incentives 0 200 400 600 800 1000 1200 1400 1600 2 4 6 8 10 12 14 16 18 20 budget degree of connection Fixed Path Sim InvSim M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 23. Outline Discovery Strategy Promotion Techniques Results Conclusions Budget Distribution Incentives 0 200 400 600 800 1000 1200 1400 1600 2 4 6 8 10 12 14 16 18 20 budget degree of connection Fixed Path Sim InvSim Incentives + Social Plasticity 0 200 400 600 800 1000 1200 1400 1600 2 4 6 8 10 12 14 16 18 20 budget degree of connection Fixed Path Sim InvSim M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 24. Outline Discovery Strategy Promotion Techniques Results Conclusions Cooperative Behavior Rate Incentives 0 200 400 600 800 1000 2 4 6 8 10 12 14 16 18 coop snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 25. Outline Discovery Strategy Promotion Techniques Results Conclusions Cooperative Behavior Rate Incentives 0 200 400 600 800 1000 2 4 6 8 10 12 14 16 18 coop snapshot Fixed Path Sim InvSim System Incentives + Social Plasticity 0 200 400 600 800 1000 2 4 6 8 10 12 14 16 18 coop snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 26. Outline Discovery Strategy Promotion Techniques Results Conclusions Success Rate Incentives 0 20 40 60 80 100 2 4 6 8 10 12 14 16 18 %successfulsearches snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 27. Outline Discovery Strategy Promotion Techniques Results Conclusions Success Rate Incentives 0 20 40 60 80 100 2 4 6 8 10 12 14 16 18 %successfulsearches snapshot Fixed Path Sim InvSim System Incentives + Social Plasticity 0 20 40 60 80 100 2 4 6 8 10 12 14 16 18 %successfulsearches snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 28. Outline Discovery Strategy Promotion Techniques Results Conclusions Path Length Incentives 0 20 40 60 80 100 2 4 6 8 10 12 14 16 18 steps snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 29. Outline Discovery Strategy Promotion Techniques Results Conclusions Path Length Incentives 0 20 40 60 80 100 2 4 6 8 10 12 14 16 18 steps snapshot Fixed Path Sim InvSim System Incentives + Social Plasticity 0 20 40 60 80 100 2 4 6 8 10 12 14 16 18 steps snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 30. Outline Discovery Strategy Promotion Techniques Results Conclusions Num. of Broken Links (Rewired) 0 200 400 600 800 1000 2 4 6 8 10 12 14 16 18 budget snapshot Fixed Path Sim InvSim System M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery
  • 31. Outline Discovery Strategy Promotion Techniques Results Conclusions Conclusions What we’ve done To combine structural changes (social plasticity) with different incentives policies in a decentralized service discovery scenario with local search. What we’ve got variable incentives work better than homogenous ones combination of mechanisms promotes cooperation in scenarios in which |nc| > |c| it increases the performance of the agents reduces the average path length increases the success rate M. Rebollo et al. (UPV) COREDEMA’13 Strategies for Cooperation Emergence in Distributed Service Discovery