Page 68 - Trust in ICT 2017
P. 68

1                                                    Trust in ICT


            •       Trustee's subjective properties, such as trustee honesty, generosity and goodness.

            •       Trustor's subjective properties, such as trustor disposition and willingness to trust.
            •       Trustor's objective properties, such as the criteria or policies specified by the trustor for a trust
                    decision.

            •       Context that the trust relationship resides in, such as the purpose of trust, the environment of trust
                    (e.g., time, location, activity, devices being used, their operational mode, etc.), and the risk of trust.

            Autonomic decision making refer to a broad interdisciplinary field interested in all aspects like economics,
            forecasting, statistical decision theory, and cognitive psychology. In general, decision making is process and
            it takes some time and effort until the choice is made, involving several activities, such as:
            •       Identification of the decision problem;
            •       Collecting and verifying relevant information;

            •       Identifying decision alternatives;
            •       Anticipating the consequences of decisions;
            •       Making the decision;
            •       Informing concerned people and public of the decision and rationale;
            •       Implementing the selected alternative;
            •       Evaluating the consequences of the decision.


























                                          Figure 17 – Autonomic control loop [57]

            There are many techniques that can be observed in the literature which address above control loop. Some
            of them are discussed below [52]:
            •       Heuristic solutions
                    This methods start from a guess about application needs and adjust this guess. Heuristic solutions
                    are  designed  for  computational  performance  or  simplicity  at  the  potential  cost  of  accuracy  or
                    precision. Such solutions generally cannot be proven to converge to the optimum or desired value.

            •       Standard control-based solutions
                    Which  employ  canonical models–  two  examples  being  discrete-time  linear models  and  discrete
                    event systems – and apply standard control techniques such as Proportional Integral controllers,
                    Proportional Integral and Derivative controllers, optimal controllers, Petri nets. Assuming the model
                    to be correct, some properties may be enforced, among which stability and convergence time are
                    probably the most important ones, thereby providing formal performance guarantees.




            60
   63   64   65   66   67   68   69   70   71   72   73