International Series of Monographs on Computer Science
1 primary work
Book 10
This book constitutes a self-contained and unified approach to automated reasoning in multiple-valued logics (MVL) developed by the author. Moreover, it contains a virtually complete account of other approaches to automated reasoning in MVL. This is the first overview of this subfield of automated reasoning ever given. Finally, a variety of applications of automated reasoning in MVL including several short case studies are listed.
Automated reasoning in non-classical logics is an essential subtask of many AI applications. Applications of MVL in particular include, for instance, hardware and software verification, reasoning with incomplete or inconsistent knowledge, and natural language processing.
Therefore, efficient theorem proving methods in MVL are essential. In the historical part of the book it is demonstrated why existing approaches are inadequate. In the original part a simple, but powerful, concept called 'sets-as-signs' is introduced in the context of semantic tableaux, and subsequently is applied to a variety of calculi including resolution and dissolution. It is shown that 'sets-as-signs' yields a many-valued extension of the well-known relationship between classical logic
and integer programming. As a consequence, automated reasoning in infinitely-valued logics can be done uniformly and efficiently for the first time.
Automated reasoning in non-classical logics is an essential subtask of many AI applications. Applications of MVL in particular include, for instance, hardware and software verification, reasoning with incomplete or inconsistent knowledge, and natural language processing.
Therefore, efficient theorem proving methods in MVL are essential. In the historical part of the book it is demonstrated why existing approaches are inadequate. In the original part a simple, but powerful, concept called 'sets-as-signs' is introduced in the context of semantic tableaux, and subsequently is applied to a variety of calculi including resolution and dissolution. It is shown that 'sets-as-signs' yields a many-valued extension of the well-known relationship between classical logic
and integer programming. As a consequence, automated reasoning in infinitely-valued logics can be done uniformly and efficiently for the first time.