Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

MAPSS, a Multi-Aspect Partner and Service Selection

281 views

Published on

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

Published in: Business, Technology
  • Be the first to comment

  • Be the first to like this

MAPSS, a Multi-Aspect Partner and Service Selection

  1. 1. 12th October 2010, Saint-Etienne, France MAPSS, a Multi-AspectPartner and Service Selection MethodZbigniew Paszkiewicz Willy Picard zpasz@kti.ue.poznan.pl picard@kti.ue.poznan.pl Poznań University of Economics
  2. 2. Agenda• Method assumptions• Social protocols• Information model• Selection method• Technical implementation• Limitations• Conclusions• Future works © UEP 2
  3. 3. Method assumptions
  4. 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 © UEP 4
  5. 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 © UEP 5
  6. 6. Social protocols
  7. 7. Social protocol• Process model encompassing Referring to VO elements, e.g. • Process structure - localization • Requirements - required set of competences » Roles - level of available resources A set of constraints on the - response time of a service » Social requirements relations among interacting - cost of the service actors (organizations and services), e.g. - past cooperation - recognition - use of service - recommendation © UEP 7
  8. 8. Abstract protocolService customer Service provider Service Process activity (organization) (organization) Partners and services are referred together as VO elements Mapping function Social network schema © UEP 8
  9. 9. Mapping Mapping social network entities social network entities to toservice description elements Social network schema entities 9
  10. 10. Multi-aspect partner and service selection (MAPSS)
  11. 11. Method information model
  12. 12. Competence description moduleService description module• Structured description of organization’s 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 © UEP 12
  13. 13. Social network module• Contents – Virtual organizations – Virtual organization partners – VOBE members – Individuals – Virtual Organization variants – Services• Basis for definition of social requirements © UEP 13
  14. 14. Indicator moduleMonitoring 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 © UEP 14
  15. 15. Selection method
  16. 16. Method outline1. Definition of VO specification2. Selection of partners and services for roles3. VO variant generation4. Performance evaluation5. VO inception In every phase, human action may lead to requirements redefinition, preference modification, repetition of a steps, and reconfiguration of used supporting tools © UEP 16
  17. 17. 1 Definition of VO specification• Set of requirements• VO planner’s preferences• VO planner’s fitness functions and acceptable requirement conformance levels © UEP 17
  18. 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 © UEP 18
  19. 19. Selection of partners and2 services for role• Set of services or organizations for each role• Sorting and filtering out elements• Requirements used – Roles – Indicators defined for roles © UEP 19
  20. 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 © UEP 20
  21. 21. 4 Performance evaluation• Fitness performance function - taking into an account various performance aspects, including – Operational performance – Effectiveness – Responsiveness – Cost• Requirements used – Performance requirements © UEP 21
  22. 22. 5 VO inception• Choosing the “winning” VO variant• Registration as a VO in – competence description – social network © UEP 22
  23. 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 © UEP 23
  24. 24. Technical implementation
  25. 25. System implementation• ErGo system• To be used in constructing sector © UEP 25
  26. 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 © UEP 26
  27. 27. 27
  28. 28. Limitations
  29. 29. Limitations• Instantiation of the whole social protocol before process launching• Single user approach• Privacy related issues are not addressed• Static method © UEP 29
  30. 30. Conclusions
  31. 31. Conclusions• Combination of social aspects, competence- based selection and performance characteristic• Already implemented• Ongoing validation in constructing sector• Source of social data © UEP 31
  32. 32. Future works
  33. 33. Future works• Continuous selection• Collaborative method• Dynamic selection method Method configurability + method instance logging = recommendation method based on process mining techniques © UEP 33
  34. 34. MAPSS, a Multi-Aspect Partner and Service Selection Method Thank youZbigniew Paszkiewicz Willy Picard zpasz@kti.ue.poznan.pl picard@kti.ue.poznan.pl Poznań University of Economics

×