The Explainer problem is to find explanations for the behaviors of systems that have been described by sets of cause-and-effect statements. Cause-and-effect statements can be used to describe a surprising large variety of different systems and situations. But it has generally been considered that this problem is only solvable by combinatorially enumerating all conceivable explanations and testing each by deduction to see if its consequences included the behavior to be described. That is, the Explainer problem has been considered to be NP-complete. If true, this would make the computing times completely impractical for all but very small problems. This NP-complete intimidation probably accounts for why this problem had not been solved long ago.
But the NP-complete problem they apparently had in mind applies to an open system where events inside the system can be caused by completely unknown events from outside. The correct Explainer problem should be closed, where only events inside the problem space can cause events inside the space. This closed problem can be solved in polynomial time, fast enough to make the Explainer quite practical. This opens up a whole new world of applications that now can be solved that previously had been believed to be unsolvable.
Read the entire article |