06.03.2013 Views

Artificial Intelligence and Soft Computing: Behavioral ... - Arteimi.info

Artificial Intelligence and Soft Computing: Behavioral ... - Arteimi.info

Artificial Intelligence and Soft Computing: Behavioral ... - Arteimi.info

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

procedure, however, no attempts have been taken to prevent the generation of<br />

absurd clauses to keep the algorithm simple.<br />

Procedure variable-instantiation-of-PRs (PRs, suspect, personmurdered);<br />

Begin<br />

For i := 1 to no-of-PRs do<br />

Begin<br />

Pick up the PR (X1, X2, ...., Xm );<br />

Form new PRs by instantiating X1, X2, ...,Xm by elements from the set of<br />

suspects <strong>and</strong> the person murdered such that X1 ≠ X2 ≠ … ≠Xm ;<br />

Record these resulting rules in a list;<br />

End For;<br />

End.<br />

Once the instantiation of the variables in the production rules is over<br />

based on the procedure presented above, the facts of the resulting rules<br />

are searched in the static database, represented in the form of a datatree.<br />

A working memory WM1 keeps track of the instantiated clauses of<br />

production rules, after these are detected in the data-tree. Moreover, if all the<br />

AND clauses of a PR are found in the data-tree, the rule is fired <strong>and</strong> the<br />

consequence of the PR is also recorded in the WM1. The <strong>info</strong>rmation stored<br />

in WM1 is subsequently used for the encoding of the place-transition pairs (of<br />

FPN), corresponding to the PR. The WM1 is then cleared for future storage.<br />

23.5.4 The Inference Engine<br />

The inference engine (IE) of the said ES comprises of five modules, namely,<br />

i) the module for searching the antecedents of PRs on the data-tree, ii) the<br />

FPN formation module, iii) belief-revision <strong>and</strong> limit-cycle detection &<br />

elimination module, iv) non-monotonic reasoning module, <strong>and</strong> v) decision<br />

making [19] <strong>and</strong> explanation tracing module.<br />

Searching antecedents of PR in the datatree: The following procedure<br />

describes the technique for searching the instantiated clauses of a PR in a<br />

data-tree.<br />

Procedure search-on-data-tree (data-tree, key-data);<br />

//searches the antecedents of PRs on data-tree//<br />

Begin<br />

Search the predicate corresponding to the key-clause in the<br />

second level of the data-tree; If search is successful Then do<br />

Begin<br />

Search the key-clause among the children of the predicate under<br />

consideration;

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

Saved successfully!

Ooh no, something went wrong!