Charles Explorer logo
🇬🇧

Discovering Equivalence Classes in Precedence Graphs

Publication at Faculty of Mathematics and Physics |
2007

Abstract

Knowledge is a specific type of information that can improve the efficiency of problem solving. During automated problem solving it may happen that knowl-edge that is known at the user level is lost in formalization.

Hence, it seems useful to re-discover this knowledge to improve the efficiency of problem solving. This paper describes a method for identifying equivalent nodes in the precedence graph with alternative branches.

Such graphs are useful for modeling manufacturing (and other) processes. Unfortunately, the formal model (which uses the precedence graph with alternatives) loses information about logical dependencies between some nodes while this information can be useful for efficient problem solving.

The paper shows that identifying equivalent nodes is an NP-hard problem. It also presents an algorithm that can discover some equivalence groups that appear frequently in real-life problems.