際際滷

際際滷Share a Scribd company logo

MAPSS, a Multi-Aspect Partner and Service Selection

Zbigniew Paszkiewicz
Zbigniew Paszkiewicz

In Service-Oriented Virtual Organization Breeding Environments (SOVOBEs), services performed by people, organizations and information systems are composed in potentially complex business processes performed by a set of partners. In a SOVOBE, the success of a virtual organization depends largely on the partner and service selection process, which determines the composition of services performed by the VO partners. In this paper requirements for a partner and service selection method for SOVOBEs are defined and a novel Multi-Aspect Partner and Service Selection method, MAPSS, is presented. The MAPSS method allows a VO planner to select appropriate services and partners based on their competences and their relations with other services/partners. The MAPSS method relies on a genetic algorithm to select the most appropriate set of partners and services. Full article: http://www.kti.ue.poznan.pl/en/node/2607

1 of 34
12th October 2010, Saint-Etienne, France MAPSS, a Multi-Aspect Partner and Service Selection Method Zbigniew Paszkiewicz Willy Picard zpasz@kti.ue.poznan.pl picard@kti.ue.poznan.pl Pozna University of Economics
Agenda Method assumptions Social protocols Information model Selection method Technical implementation Limitations Conclusions Future works 息 UEP 2
Method assumptions
Current approaches Missing an approach integrating Competence-based approach Performance evaluation Social aspects Separation of service search from the selection of partners Not mature concepts of modeling social requirements 息 UEP 4
MAPSS method assumptions Partner and service selection method supporting Social aspects Competence-based selection Partner and network performance characteristic Human control over the process The novelty of the proposed method lays in the Multi-variant analysis of the concepts combination Requirement-based approach 息 UEP 5
Social protocols
Social protocol Process model encompassing Referring to VO elements, e.g. Process structure - localization Requirements - required set of competences 損 Roles - level of available resources A set of constraints on the - response time of a service 損 Social requirements relations among interacting - cost of the service actors (organizations and services), e.g. - past cooperation - recognition - use of service - recommendation 息 UEP 7
Abstract protocol Service customer Service provider Service Process activity (organization) (organization) Partners and services are referred together as VO elements Mapping function Social network schema 息 UEP 8
Mapping Mapping social network entities social network entities to to service description elements Social network schema entities 9
Multi-aspect partner and service selection (MAPSS)
Method information model
Competence description module Service description module Structured description of organizations competences and services Basis for definition of roles Organization/service search based on submitted criteria Evaluation of the conformance of an organization to a set of requirements In the current implementation a modified 4-C model (Ermilova, Afsarmanesh) is used 息 UEP 12
Social network module Contents Virtual organizations Virtual organization partners VOBE members Individuals Virtual Organization variants Services Basis for definition of social requirements 息 UEP 13
Indicator module Monitoring module Not only performance indicators Definition of complex requirements involving various aspects in a single indicator, e.g. competences and social relations Monitoring of requirement values and notification of changes 息 UEP 14
Selection method
Method outline 1. Definition of VO specification 2. Selection of partners and services for roles 3. VO variant generation 4. Performance evaluation 5. VO inception In every phase, human action may lead to requirements redefinition, preference modification, repetition of a steps, and reconfiguration of used supporting tools 息 UEP 16
1 Definition of VO specification Set of requirements VO planners preferences VO planners fitness functions and acceptable requirement conformance levels 息 UEP 17
1 Definition of VO specification Requirement types Roles Defined in abstract social Social requirements protocol Indicators i.e. performance requirements Aspects VO elements User defined Process Subsets of partners Subsets of services 息 UEP 18
Selection of partners and 2 services for role Set of services or organizations for each role Sorting and filtering out elements Requirements used Roles Indicators defined for roles 息 UEP 19
3 Generation of VO variants Genetic algorithm - VO Variant - determination of the best fitted possible VO variants composition of Fitness function - estimation of partners and the level of satisfaction of social requirements services Threshold value - used to filter out the VO variants Requirements used Social requirements 息 UEP 20
4 Performance evaluation Fitness performance function - taking into an account various performance aspects, including Operational performance Effectiveness Responsiveness Cost Requirements used Performance requirements 息 UEP 21
5 VO inception Choosing the winning VO variant Registration as a VO in competence description social network 息 UEP 22
VO specification structure Aspects Roles Social req. Performance req. Partner Phase 2 - - Service Phase 2 - - Subset of partners - Phase 3 - Subset of services - Phase 3 Phase 4 Process - - Phase 4 Phase 2. Selection of partners and services for roles Phase 3. VO variant generation Phase 4. Performance evaluation 息 UEP 23
Technical implementation
System implementation ErGo system To be used in constructing sector 息 UEP 25
Implementation environment Programming platform Java Web application Google web Toolkit System modularity OSGi Genetic algorithm Java Genetic Algorithm Package Social network Java Universal Network Graph Framework 息 UEP 26
27
Limitations
Limitations Instantiation of the whole social protocol before process launching Single user approach Privacy related issues are not addressed Static method 息 UEP 29
Conclusions
Conclusions Combination of social aspects, competence- based selection and performance characteristic Already implemented Ongoing validation in constructing sector Source of social data 息 UEP 31
Future works
Future works Continuous selection Collaborative method Dynamic selection method Method configurability + method instance logging = recommendation method based on process mining techniques 息 UEP 33
MAPSS, a Multi-Aspect Partner and Service Selection Method Thank you Zbigniew Paszkiewicz Willy Picard zpasz@kti.ue.poznan.pl picard@kti.ue.poznan.pl Pozna University of Economics

More Related Content

MAPSS, a Multi-Aspect Partner and Service Selection

  • 1. 12th October 2010, Saint-Etienne, France MAPSS, a Multi-Aspect Partner and Service Selection Method Zbigniew Paszkiewicz Willy Picard zpasz@kti.ue.poznan.pl picard@kti.ue.poznan.pl Pozna University of Economics
  • 2. Agenda Method assumptions Social protocols Information model Selection method Technical implementation Limitations Conclusions Future works 息 UEP 2
  • 4. Current approaches Missing an approach integrating Competence-based approach Performance evaluation Social aspects Separation of service search from the selection of partners Not mature concepts of modeling social requirements 息 UEP 4
  • 5. MAPSS method assumptions Partner and service selection method supporting Social aspects Competence-based selection Partner and network performance characteristic Human control over the process The novelty of the proposed method lays in the Multi-variant analysis of the concepts combination Requirement-based approach 息 UEP 5
  • 7. Social protocol Process model encompassing Referring to VO elements, e.g. Process structure - localization Requirements - required set of competences 損 Roles - level of available resources A set of constraints on the - response time of a service 損 Social requirements relations among interacting - cost of the service actors (organizations and services), e.g. - past cooperation - recognition - use of service - recommendation 息 UEP 7
  • 8. Abstract protocol Service customer Service provider Service Process activity (organization) (organization) Partners and services are referred together as VO elements Mapping function Social network schema 息 UEP 8
  • 9. Mapping Mapping social network entities social network entities to to service description elements Social network schema entities 9
  • 10. Multi-aspect partner and service selection (MAPSS)
  • 12. Competence description module Service description module Structured description of organizations competences and services Basis for definition of roles Organization/service search based on submitted criteria Evaluation of the conformance of an organization to a set of requirements In the current implementation a modified 4-C model (Ermilova, Afsarmanesh) is used 息 UEP 12
  • 13. Social network module Contents Virtual organizations Virtual organization partners VOBE members Individuals Virtual Organization variants Services Basis for definition of social requirements 息 UEP 13
  • 14. Indicator module Monitoring module Not only performance indicators Definition of complex requirements involving various aspects in a single indicator, e.g. competences and social relations Monitoring of requirement values and notification of changes 息 UEP 14
  • 16. Method outline 1. Definition of VO specification 2. Selection of partners and services for roles 3. VO variant generation 4. Performance evaluation 5. VO inception In every phase, human action may lead to requirements redefinition, preference modification, repetition of a steps, and reconfiguration of used supporting tools 息 UEP 16
  • 17. 1 Definition of VO specification Set of requirements VO planners preferences VO planners fitness functions and acceptable requirement conformance levels 息 UEP 17
  • 18. 1 Definition of VO specification Requirement types Roles Defined in abstract social Social requirements protocol Indicators i.e. performance requirements Aspects VO elements User defined Process Subsets of partners Subsets of services 息 UEP 18
  • 19. Selection of partners and 2 services for role Set of services or organizations for each role Sorting and filtering out elements Requirements used Roles Indicators defined for roles 息 UEP 19
  • 20. 3 Generation of VO variants Genetic algorithm - VO Variant - determination of the best fitted possible VO variants composition of Fitness function - estimation of partners and the level of satisfaction of social requirements services Threshold value - used to filter out the VO variants Requirements used Social requirements 息 UEP 20
  • 21. 4 Performance evaluation Fitness performance function - taking into an account various performance aspects, including Operational performance Effectiveness Responsiveness Cost Requirements used Performance requirements 息 UEP 21
  • 22. 5 VO inception Choosing the winning VO variant Registration as a VO in competence description social network 息 UEP 22
  • 23. VO specification structure Aspects Roles Social req. Performance req. Partner Phase 2 - - Service Phase 2 - - Subset of partners - Phase 3 - Subset of services - Phase 3 Phase 4 Process - - Phase 4 Phase 2. Selection of partners and services for roles Phase 3. VO variant generation Phase 4. Performance evaluation 息 UEP 23
  • 25. System implementation ErGo system To be used in constructing sector 息 UEP 25
  • 26. Implementation environment Programming platform Java Web application Google web Toolkit System modularity OSGi Genetic algorithm Java Genetic Algorithm Package Social network Java Universal Network Graph Framework 息 UEP 26
  • 27. 27
  • 29. Limitations Instantiation of the whole social protocol before process launching Single user approach Privacy related issues are not addressed Static method 息 UEP 29
  • 31. Conclusions Combination of social aspects, competence- based selection and performance characteristic Already implemented Ongoing validation in constructing sector Source of social data 息 UEP 31
  • 33. Future works Continuous selection Collaborative method Dynamic selection method Method configurability + method instance logging = recommendation method based on process mining techniques 息 UEP 33
  • 34. MAPSS, a Multi-Aspect Partner and Service Selection Method Thank you Zbigniew Paszkiewicz Willy Picard zpasz@kti.ue.poznan.pl picard@kti.ue.poznan.pl Pozna University of Economics

Editor's Notes

  1. WP + ZP