SDML can be used to model the past and present to predict the future and determine what to do to achieve your desired outcome. The science fiction author, Isaac Asimov coined the term "psychohistory" for this in his brilliant Foundation series, but the same sort of thing is done by people's brains for everyday sub-tasks.

If you want to learn SDML, I recommend learning Prolog first because there is much more documentation for that language and the backward chaining aspects of SDML are based on Prolog without the imperative features that undermine Prolog's relationship with logic. In particular, the "cut" (an exclamation mark) is used in Prolog for reasons of efficiency and much of my work was devoted to coming up with declarative ways of achieving efficiency, for both forward and backward chaining, without resorting to such measures.

To access a page containing recommendations for introductions to Prolog, click here. I think I learnt Prolog via an earlier edition of Clocksin and Mellish's "Programming in Prolog".