Skip to Main content Skip to Navigation
Theses

Modal memory logics

Abstract : From ancient times to the present day, the field of logic has gained significant strength and now it actively contributes to many different areas, such as philos- ophy, mathematics, linguistic, computer science, artificial intelligence, hardware manufacture, etc. Each of these scenarios has specific needs, that range from very concrete requirements, like an efficient inference method, to more abstract theoretical properties, like a neat axiomatic system. Given this wide diversity of uses, a motley collection of formal languages has been developed. For many years, classical languages (mainly classical first order logic) were the alternative, but this assortment of applications made other types of logics also attractive in many situations. Imagine that the time for choosing a logic for some specific task arrives. How can we decide which is the one that fits best? Which properties should we look for? How can we ?measure? a logic with respect to others? These are not easy questions, and there is not a general recipe one can follow. In this thesis we are just going to restrict these questions to a particular family of logics, and in that context we will investigate theoretical aspects that help to answer some of these concerns. Much can be discovered by carefully analyzing appealing cases, and our contribution will be developed having that philosophy in mind. Propositional modal logics offer an alternative to traditional languages. They can be regarded as a set of tools that allow to design logics specially tailored for specific tasks, having a fine-grained control on their expressivity. Additionally, modal logics turned out to have a good computational behavior, which proved to be quite robust under extensions. These characteristics, among others, placed modal logics as an attractive alternative to classical languages. In this dissertation we are going to present a new family of modal logics called memory logics. Traditional modal logics enables to describe relational structures from a local perspective. But what about changing the structure? We want to explore the addition of an explicit storage structure to modal logics, a mem- ory, that allows to model dynamic behavior through explicit memory operators. These operators store or retrieve information to and from the memory. Natu- rally, depending on which type of storage structure we want, and which memory operators are available, the resulting logic will enjoy different properties that are worth investigating. The thesis is organized as follows. In Chapter 1 we start by giving a brief recap of how modal logic was born, showing the different historical perspectives used to look at modal logic. Then we formally present the basic modal logic and a set of extended operators that helps grasp the modal ?flavor? of some richer languages. We finish this chapter by giving a first glance of memory logics, and showing how they can help to model state when we choose to use a set as storage structure. Chapter 2 is devoted to present memory logics in detail. We show some examples that can be described by adding a set to standard relational structures, and the usual set operators to add elements and test membership. We then show some other memory operators that can be considered, and we discuss the possibility of adding constraints to the interplay between memory and modal operators. These constraints can be regarded as a way to have a finer-grained control on the logic expressivity. Since we have made changes to classical modal logics, we are interested in analyzing the impact those changes cause in the resulting logics. Therefore, the rest of this chapter presents a basic logic toolkit through which we can analyze this new family of logics. This toolkit can be seen as an outline that organizes the rest of the thesis and that allows to analyze memory logics in terms of expressivity, complexity, interpolation and proof theory. The rest of the chapters investigate each of these aspects in detail. In Chap- ters 3 and 4 we explore the expressive power of several memory logics and we study the decidability of their satisfiability problem. In the decidable cases, we determine their computational complexity. We analyze the impact of the differ- ent memory operators we consider, and how they interact. We also study other memory containers, such as a stack. Then, in Chapter 5, we analyze Craig inter- polation and Beth definability for some memory logic fragments. We also study memory logics from a proof theoretical perspective. In Chapter 6 and 7 we turn to Hilbert style axiomatizations and tableau systems, and we characterize several fragments of the memory logic family mostly using techniques borrowed from hy- brid logics. We close in Chapter 8 with some concluding remarks, open problems and directions for further research.
Document type :
Theses
Complete list of metadata

Cited literature [90 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01748219
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 11:28:51 AM
Last modification on : Tuesday, April 24, 2018 - 1:53:23 PM
Long-term archiving on: : Friday, September 14, 2018 - 12:51:39 AM

File

SCD_T_2009_0130_MERA.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01748219, version 1

Collections

Citation

Sergio Fernando Mera. Modal memory logics. Other [cs.OH]. Université Henri Poincaré - Nancy 1, 2009. English. ⟨NNT : 2009NAN10130⟩. ⟨tel-01748219⟩

Share

Metrics

Record views

72

Files downloads

158