23.02.2015 Views

D.3.3 ALGORITHMS FOR INCREMENTAL ... - SecureChange

D.3.3 ALGORITHMS FOR INCREMENTAL ... - SecureChange

D.3.3 ALGORITHMS FOR INCREMENTAL ... - SecureChange

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Dealing with Known Unknowns: A Goal-based Approach 7<br />

troduced, or objectives’ status is changed e.g., from ‘optional’ to ‘compulsory’,<br />

or from ‘compulsory’ to ‘obsoleted’.<br />

In an evolution-aware developing process, the selection of optimal design alternative(s)<br />

for the implementation of an enterprise system should carefully take<br />

into account observable evolutions, otherwise it may be incapable to support the<br />

enterprise objectives in future.<br />

To support decision making in dealing with the evolution of enterprise objectives,<br />

these two kinds of evolutions are incorporated into enterprise models in<br />

terms of evolution rules. The basic idea of evolution rules is to capture the snapshots<br />

of enterprise objectives before and after evolutions. Evolutions rules are formally<br />

defined as follows:<br />

Definition 2 (Controllable rule) A controllable rule r c is a set of tuples 〈EM, EM i 〉<br />

that consists of an original model EM and its possible design alternative EM i .<br />

r c = [ o<br />

nEM −→ ∗ EM i<br />

i<br />

Example 1 (Controllable rule) Consider again Fig. 1, the objective g 5 −Data exchanged<br />

is refined into only one lower level objective g 9 −Data exchanged with adjacent sectors.<br />

However an organization might chose to integrate the DMAN in the process and<br />

thus might choose between two other goals g 12 −Basic data exchanged with DMAN and<br />

g 13 −Advanced data exchanged with DMAN.<br />

Definition 3 (Observable rule) An observable rule r o is a set of triples 〈EM, p i , EM i 〉<br />

that consists of an original model EM and its potential evolution EM i . The probability<br />

that EM evolves to EM i is p i . All these probabilities should sum up to one.<br />

r o = [ n<br />

EM p o<br />

i<br />

−→ EM i<br />

i<br />

Example 2 (Observable rule) Consider again Fig. 1, the objective g 5 −Data exchanged<br />

is refined into only one lower level objective g 9 −Data exchanged with adjacent sectors.<br />

One potential evolution for is that a regulatory body requires that the organization<br />

should also achieve both goal g 12 −Basic data exchanged with DMAN and goal g 13 −Advanced<br />

data exchanged with DMAN albeit for different functionalities. If this possibility actually<br />

happens the organization should meet g 9 , g 12 , g 13 in order to meet g 5 . However, the<br />

discussion in the regulatory body is still quite open and another option might be<br />

to actually leave partners to chose whether to impose g 9 −Data exchanged with adjacent<br />

sectors and g 12 −Basic data exchanged with DMAN and leave operators the choice of implementing<br />

g 9 and g 13 −Advanced data exchanged with DMAN. This latter option, which combines<br />

the one from Example2, is denoted by EM 1 as it is the most likely outcome,<br />

the former we denote it by EM 2 . Obviously, there is also another possibility: EM<br />

might not evolve and therefore remain unchanged (decision of regulatory body<br />

is post-poned after first deployment trials of DMAN). These possibilities exclusively<br />

happen with some uncertainties. Thus, the observable rule is would be the<br />

following one.<br />

j<br />

ff<br />

r o1 = EM p 1<br />

−−→ EM 1 , EM p 2<br />

−−→ EM 2 , EM 1−p 1 −p 2<br />

−−−−−−→ EM

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!