Algorithms for the Forgetting-Based Inconsistency Measure[go to overview]
Knowledge bases are sets of propositional logic formulas and can be either consistent or inconsistent. Knowledge bases can be inconsistent to varying degrees: Given two inconsistent knowledge bases, one might only have a few contradictions while the other has many. So called inconsistency measures quantify these different degrees of inconsistency. There are many different inconsistency measures based on different ideas on what consitutes the severeness of inconsisteny. My bachelor’s thesis aimed to find algorithms to compute one such inconsistency measure: The Forgetting-Based Inconsistency Measure. In the first half of the presentation I will explain what it means for a knowledge base to be consistent or inconsistent and how the “forgetting” operation can be used to quanitfy the degree of inconsistency of a given knowledge base. In the latter half I will explain the basic idea behind the binary search algorithm I implemented to compute The Forgetting-Based Inconsistency Measure. The SAT-Reduction the algorithm employs will be covered as well. Finally I will summarize my findings on how the algorithm performs on different sample sets of knowledge bases and explain the limitations of the chosen computation method.
12.11.20 - 10:15
via Big Blue Button