Page 80 - Trust in ICT 2017
P. 80

1                                                    Trust in ICT


                    The benefit of this approach is that it permits the use of explicitly stated expectations, such as
                    contract clauses, in the decision about whether to trust. This approach needs not to be limited to
                    contracts; social norms can equally be considered. With this in mind, this type of approach may be
                    suitable for reasoning explicitly about the integrity of agents, as well as their competence, based on
                    past performance with respect to norms and contracts.
                    The  model  proposed  by  Smith  and  des  Jardins  addresses  the  decision  problem  of  agents  by
                    modelling interactions as Iterated Prisoner’s Dilemma games. These are a repeated variant of the
                    classic  Prisoner’s  Dilemma  game  (Axelrod  and  Hamilton),  where  the  ‘players’  have  a  personal
                    incentive to behave in an untrustworthy way.

            8.3.3   Implement a trust reasoner using rule languages [83]
            Ontologies  are  formal  definitions  of  concepts  and  the  relationships  between  them.  The  Web  Ontology
            Language OWL 2 is a W3C Recommendation since 2009. It is based on Description Logics (DLs), a family of
            knowledge representation formalisms. OWL 2 RL (Rule Language) reasoning systems allow for rule-based
            reasoning.  OWL  2  Query  Language  (QL)  supports  conjunctive  query  answering  against  large  volumes  of
            instance data that is stored in relational database systems. OWL 2 EL aims at applications that employ large
            ontologies.
            A reasoner is a program that infers logical consequences from a set of explicitly asserted facts or axioms and
            typically provides automated support for reasoning tasks such as classification, debugging and querying. For
            OWL 2 EL, scalable implementations of dedicated reasoning algorithms are available. A question is whether
            these implementations perform better on OWL 2 EL ontologies than traditional reasoning engines, which
            have  been  designed  for  much  more  expressive  languages.  Sematic  tableau  algorithms  can  be  highly
            optimized, so that they are not necessarily outperformed by straightforward implementations of polynomial-
            time algorithms.

            Here are some prospective reasoners that we can use for trust.
            CB (Consequence-based reasoner, University of Oxford) is an implementation of a reasoning procedure for
            Horn Ontologies, i.e. SHIQ ontologies that can be translated to the Horn fragment of first-order logic. CB’s
            reasoning procedure can be regarded as an extension of the completion-based procedure for EL++ ontologies
            and works by deriving new consequent axioms. It is theoretically optimal for Horn SHIQ ontologies as well as
            for the common fragment of EL++ and SHIQ.
            FaCT++ (Fast Classification of Terminologies, University of Manchester) is the new generation of the OWL
            DL reasoner FaCT. It supports OWL DL and a subset of OWL 2 that is more expressive than the ontologies in
            other ontologies. FaCT++ is implemented in C++ and based on optimized tableaux algorithms.
            HermiT (University of Oxford) can determine whether or not a given ontology is consistent and identify
            subsumption relationships between concepts, among other features. HermiT is based on a “hypertableau”
            calculus.

            TrOWL (Tractable reasoning infrastructure for OWL 2, University of Aberdeen) is the common interface to
            a number of reasoners. TrOWL Quill provides reasoning services over OWL 2 QL. TrOWL REL is an optimized
            implementation  of  the  CEL  algorithm  that  provides  reasoning  over  OWL  2  EL.  It  employs  a  syntactic
            approximation from OWL 2 DL to OWL 2 EL to enable OWL 2 DL ontologies to be classified within polynomial
            time  [41].  This  approximation  is  soundness-preserving  but  sacrifices  completeness.  To  support  full  DL
            reasoning, TrOWL allows for the use of heavyweight plugin reasoners, such as FaCT++, Pellet, HermiT and
            RacerPro.

            8.4     A reputation and knowledge based trust model and decision making mechanism

            There are numerous trust solutions have been proposed for each environment (e.g. P2P, MAS, e-commerce,
            etc.), in this section, it aims at developing a trust service platform that cooperates with applications and
            services to for the trust in future social IoT environments.






            72
   75   76   77   78   79   80   81   82   83   84   85