Use this URL to cite or link to this record in EThOS:
Title: Decentralised group formation in pervasive environments
Author: De Silva, Hasini
ISNI:       0000 0004 2718 0400
Awarding Body: University of Surrey
Current Institution: University of Surrey
Date of Award: 2011
Availability of Full Text:
Access from EThOS:
Access from Institution:
Group collaborations involve an amalgamation of entities capable of achieving related goals. Facilitating such collaborations in pervasive environments requires automated formation and maintenance of groups consisting of individuals collectively possessing characteristics required to achieve the collaboration goals. Most existing group formation solutions are application-specific in which the nature of the formed groups are monotonic in terms of the composition of the group. Furthermore, most of these solutions are fixed in terms of the nature of member attributes considered as the basis for grouping. The group formation process becomes especially challenging due to the decentralised nature of the pervasive environments, since no single central entity capable of coordinating the process exists, which has a global view of all potential members. This necessitates the cooperation of potential members when forming groups. This thesis therefore proposes a generic and decentralised solution for automating group formation and maintenance in a pervasive environment, regardless of the nature or the requirements of the environment. A structure for defining group formation criteria based on features of potential members is proposed in this work that does not impose restrictions on the nature of the formed groups. Generic group formation approaches are proposed that successfully form decentralised groups based on the group formation criteria, regardless of the domain of the application and are superior in performance to the closest approach found in the literature. Group maintenance solutions are proposed that enable groups to withstand dynamicity, which is addressed in terms of the notion of events. This includes feature updates of individuals, in addition to appearance and disappearance of individuals, which are usually tackled in the literature. The generality of the solutions is facilitated by introducing a feature description scheme, which contains the methods of knowledge derivation related to each feature of concern.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available