Finitary proof systems for Kozen s µ

Σχετικά έγγραφα
Sequent Calculi for the Modal µ-calculus over S5. Luca Alberucci, University of Berne. Logic Colloquium Berne, July 4th 2008

Lecture 2. Soundness and completeness of propositional logic

Ordinal Arithmetic: Addition, Multiplication, Exponentiation and Limit

Every set of first-order formulas is equivalent to an independent set

LTL to Buchi. Overview. Buchi Model Checking LTL Translating LTL into Buchi. Ralf Huuck. Buchi Automata. Example

Nowhere-zero flows Let be a digraph, Abelian group. A Γ-circulation in is a mapping : such that, where, and : tail in X, head in

Example Sheet 3 Solutions

EE512: Error Control Coding

C.S. 430 Assignment 6, Sample Solutions

2 Composition. Invertible Mappings

Abstract Storage Devices

A Lambda Model Characterizing Computational Behaviours of Terms

Overview. Transition Semantics. Configurations and the transition relation. Executions and computation

From the finite to the transfinite: Λµ-terms and streams

The Simply Typed Lambda Calculus

ω ω ω ω ω ω+2 ω ω+2 + ω ω ω ω+2 + ω ω+1 ω ω+2 2 ω ω ω ω ω ω ω ω+1 ω ω2 ω ω2 + ω ω ω2 + ω ω ω ω2 + ω ω+1 ω ω2 + ω ω+1 + ω ω ω ω2 + ω

Jesse Maassen and Mark Lundstrom Purdue University November 25, 2013

Fractional Colorings and Zykov Products of graphs

Fourier Series. MATH 211, Calculus II. J. Robert Buchanan. Spring Department of Mathematics

CRASH COURSE IN PRECALCULUS

Commutative Monoids in Intuitionistic Fuzzy Sets

Reminders: linear functions

4.6 Autoregressive Moving Average Model ARMA(1,1)

3.4 SUM AND DIFFERENCE FORMULAS. NOTE: cos(α+β) cos α + cos β cos(α-β) cos α -cos β

The λ-calculus. Lecturer: John Wickerson. Phil Wadler

Arithmetical applications of lagrangian interpolation. Tanguy Rivoal. Institut Fourier CNRS and Université de Grenoble 1

Math 446 Homework 3 Solutions. (1). (i): Reverse triangle inequality for metrics: Let (X, d) be a metric space and let x, y, z X.

Homomorphism in Intuitionistic Fuzzy Automata

Other Test Constructions: Likelihood Ratio & Bayes Tests

A Note on Intuitionistic Fuzzy. Equivalence Relation

MINIMAL CLOSED SETS AND MAXIMAL CLOSED SETS

Lecture 2: Dirac notation and a review of linear algebra Read Sakurai chapter 1, Baym chatper 3

The circle theorem and related theorems for Gauss-type quadrature rules

Bounding Nonsplitting Enumeration Degrees

ΚΥΠΡΙΑΚΗ ΕΤΑΙΡΕΙΑ ΠΛΗΡΟΦΟΡΙΚΗΣ CYPRUS COMPUTER SOCIETY ΠΑΓΚΥΠΡΙΟΣ ΜΑΘΗΤΙΚΟΣ ΔΙΑΓΩΝΙΣΜΟΣ ΠΛΗΡΟΦΟΡΙΚΗΣ 19/5/2007

The challenges of non-stable predicates

ST5224: Advanced Statistical Theory II

Finite Field Problems: Solutions

Statistical Inference I Locally most powerful tests

Chapter 3: Ordinal Numbers

Uniform Convergence of Fourier Series Michael Taylor

Areas and Lengths in Polar Coordinates

Affine Weyl Groups. Gabriele Nebe. Summerschool GRK 1632, September Lehrstuhl D für Mathematik

Solutions to Exercise Sheet 5

Models for Probabilistic Programs with an Adversary

On a four-dimensional hyperbolic manifold with finite volume

About these lecture notes. Simply Typed λ-calculus. Types

Approximation of distance between locations on earth given by latitude and longitude

Inverse trigonometric functions & General Solution of Trigonometric Equations

Some new generalized topologies via hereditary classes. Key Words:hereditary generalized topological space, A κ(h,µ)-sets, κµ -topology.

Chap. 6 Pushdown Automata

Generating Set of the Complete Semigroups of Binary Relations

Areas and Lengths in Polar Coordinates

5. Choice under Uncertainty

HOMEWORK 4 = G. In order to plot the stress versus the stretch we define a normalized stretch:

ANSWERSHEET (TOPIC = DIFFERENTIAL CALCULUS) COLLECTION #2. h 0 h h 0 h h 0 ( ) g k = g 0 + g 1 + g g 2009 =?

Homework 8 Model Solution Section

λρ-calculus 1. each λ-variable is a λρ-term, called an atom or atomic term; 2. if M and N are λρ-term then (MN) is a λρ-term called an application;

Homework 3 Solutions

Lecture 13 - Root Space Decomposition II

ORDINAL ARITHMETIC JULIAN J. SCHLÖDER

Matrices and Determinants

2. THEORY OF EQUATIONS. PREVIOUS EAMCET Bits.

Chapter 6: Systems of Linear Differential. be continuous functions on the interval

Phys460.nb Solution for the t-dependent Schrodinger s equation How did we find the solution? (not required)

Section 8.3 Trigonometric Equations

SOLVING CUBICS AND QUARTICS BY RADICALS

LAP 2013 Problems in formulating the consecution calculus of contraction less relevant logics

Verification. Lecture 12. Martin Zimmermann

2. Let H 1 and H 2 be Hilbert spaces and let T : H 1 H 2 be a bounded linear operator. Prove that [T (H 1 )] = N (T ). (6p)

Parametrized Surfaces

Econ 2110: Fall 2008 Suggested Solutions to Problem Set 8 questions or comments to Dan Fetter 1

Quadratic Expressions

DiracDelta. Notations. Primary definition. Specific values. General characteristics. Traditional name. Traditional notation

An Inventory of Continuous Distributions

Physical DB Design. B-Trees Index files can become quite large for large main files Indices on index files are possible.

Simulation of G i with prenex cuts

Αλγόριθμοι και πολυπλοκότητα NP-Completeness (2)

The Probabilistic Method - Probabilistic Techniques. Lecture 7: The Janson Inequality

ΚΥΠΡΙΑΚΟΣ ΣΥΝΔΕΣΜΟΣ ΠΛΗΡΟΦΟΡΙΚΗΣ CYPRUS COMPUTER SOCIETY 21 ος ΠΑΓΚΥΠΡΙΟΣ ΜΑΘΗΤΙΚΟΣ ΔΙΑΓΩΝΙΣΜΟΣ ΠΛΗΡΟΦΟΡΙΚΗΣ Δεύτερος Γύρος - 30 Μαρτίου 2011

On the Galois Group of Linear Difference-Differential Equations

1. Introduction and Preliminaries.

k A = [k, k]( )[a 1, a 2 ] = [ka 1,ka 2 ] 4For the division of two intervals of confidence in R +

PARTIAL NOTES for 6.1 Trigonometric Identities

Second Order Partial Differential Equations

Homomorphism of Intuitionistic Fuzzy Groups

derivation of the Laplacian from rectangular to spherical coordinates

Partial Differential Equations in Biology The boundary element method. March 26, 2013

6.1. Dirac Equation. Hamiltonian. Dirac Eq.

Durbin-Levinson recursive method

Main source: "Discrete-time systems and computer control" by Α. ΣΚΟΔΡΑΣ ΨΗΦΙΑΚΟΣ ΕΛΕΓΧΟΣ ΔΙΑΛΕΞΗ 4 ΔΙΑΦΑΝΕΙΑ 1

Lecture 15 - Root System Axiomatics

Second Order RLC Filters

Distances in Sierpiński Triangle Graphs

Simulating non-prenex cuts in quantified propositional calculus

Local Approximation with Kernels

Proving with Computer Assistance Lecture 2. Herman Geuvers

Heisenberg Uniqueness pairs

Iterated trilinear fourier integrals with arbitrary symbols

The semiclassical Garding inequality

Transcript:

Finitary proof systems for Kozen s µ Bahareh Afshari Graham Leigh TU Wien University of Gothenburg homc & cdps 16, Singapore 1 / 17

Modal µ-calculus Syntax: p p φ ψ φ ψ φ φ x µx φ νx φ Semantics: For Kripke structure K = (W,, λ) and valuation V : Var 2 W p V K = {u W p λ(u)} φ ψ K V = φ K V ψ K V φ V K = {u W v(u v v φ K V )} x K V = V (x) and similarly for p, and µx φ V K = least fixpoint of the function X φ K V [x X ] { = X W φ K V [x X ] X } νx φ V K = greateast fixpoint of the function X φ K V [x X ] { } = X W X φ K V [x X ] Examples: µx( x p); νx( x p); νxµy( y (p x)) Duality: Define φ as the De Morgan dual of φ: µxφ(x) = νxφ(x) φ V K = W \ φ K y V (y) 2 / 17

Validity and proofs Let φ be a closed formula Define K = φ iff φ K = W where K = (W,, λ), = φ (φ is valid) iff K = φ for every K Theorem (Kozen 1983; Walukiewicz 2000) For every closed φ, = φ iff Koz φ Soundness: Proved by Kozen (1983) Completeness: Kozen: aconjuntive fragment; Walukiewicz: full µ-calculus: 1 Completeness of disjunctive fragment: tableaux 2 Provable equivalence between disjunctive and µ-formulæ: tableaux, games/automata 3 / 17

almost always implies infinitely often Consider the formula ψ = µxνyφ(x, y) ν yµxφ(x, y) ψ is valid easy semantic argument Koz ψ non-trivial Questions: 1 Is there a more direct/constructive proof of completeness? 2 Is cut necessary? 3 Are there other natural sound and complete finitary proof systems? 4 / 17

Tableaux proofs A tableau is a Fix-tree in which every infinite path contains a ν-thread Theorem (Niwinski, Walukiewicz 1996; Studer 2008; Friedman 2013) For every closed guardedguarded formula φ, = φ iff there exists a tableau for φ ( ) Y (νxy ), X (νyx ) νx νxy, X (νyx ) Y (νxy ),νyx ( ) φ(νxy,y (νxy )), φ(x (νyx ),νyx ) µy + µ x Y (νxy ), X (νyx ) ( ) νy Y (νxy ),νyx ( ) νx νxµyφ,νyµxφ where Y (x) = µyφ(x, y) and X (y) = µxφ(x, y) 5 / 17

Stirling s tableaux proofs with names Fix a set of names for each variable: N x = {x 0, x 1, } and N = x:var N x An annotated sequent is an expression a 0 φ a 1 1,, φa n n st a 0,, a n N Definition A Stirling proof is a finite tree built from rules Fix N + reset x + ν x + exp st 1 For every sequent a φ a 0 0,, φa k and every i k, a k i a; 2 Every non-axiom leaf has the configuration axb axb Π resetx axb Π axb where: 21 x appears in every sequent between repetition 22 reset x occurs between the two nodes We write Stir Γ if there exists a Stirling proof with root ε {φ ε φ Γ} Theorem (Stirling 2014) Tableaux can be effectively transformed into Stirling proofs 6 / 17

Stirling proofs: interpreting tableaux [xy Y (νxy ) x, X (νyx ) y ] reset x [xy Y (νxy ) x, X (νyx ) y ] reset y xyx Y (νxy ) xx, X (νyx ) y ν x xy νxy x, X (νyx ) y xyy Y (νxy ) x, X (νyx ) yy ν y xy Y (νxy ) x,νyx y xy φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ x + µ y xy Y (νxy ) x, X (νyx ) y ( ) νy x Y (νxy ) x,νyx ε ν x ε νxy ε,νyx ε Pros: 1 Fintary proofs 2 Proofs constructed semantically Cons: 1 Non-locality 2 Guessing resets 7 / 17

Circular proofs with ν-closure Consider the rule Definition [ Γ,νxφ(x) ax ] Γ, φ(νxφ) ax where a x, x is not in Γ or a clo Γ,νxφ a Clo Γ iff there exists a finite tree built from rules Fix N + ν x + clo satisfying 1 Every sequent has the form ε φ a 1 1,, φa k k ; 2 Every leaf is either an axiom or discharged by an application of clo: Theorem Stir Γ implies Clo Γ Proof 1 Unravel Stirling proofs; 2 Delete resets; 3 Search for clo 8 / 17

Example: Stirling to circular proofs (I) [xy Y (νxy ) x, X (νyx ) y ] reset x [xy Y (νxy ) x, X (νyx ) y ] reset y xyx Y (νxy ) xx, X (νyx ) y ν x xy νxy x, X (νyx ) y xyy Y (νxy ) x, X (νyx ) yy ν y xy Y (νxy ) x,νyx y xy φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ x + µ y xy Y (νxy ) x, X (νyx ) y ( ) νy x Y (νxy ) x,νyx ε ν x ε νxy ε,νyx ε 9 / 17

Example: Stirling to circular proofs (II) Y (νxy ) x, X (νyx ) y reset x Y (νxy ) x, X (νyx ) y reset y Y (νxy ) xx, X (νyx ) y ν x νxy x, X (νyx ) y Y (νxy ) x, X (νyx ) yy ν y Y (νxy ) x,νyx y φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ Y (νxy ) x, X (νyx ) y reset x Y (νxy ) x, X (νyx ) y reset y Y (νxy ) xx, X (νyx ) y ν x νxy x, X (νyx ) y Y (νxy ) x, X (νyx ) yy ν y Y (νxy ) x,νyx y φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ x + µ y Y (νxy ) x, X (νyx ) y ν y Y (νxy ) x,νyx ε ν x νxy ε,νyx ε 10 / 17

Example: Stirling to circular proofs (III) Y (νxy ) xx x, X (νyx ) y reset x Y (νxy ) xx, X (νyx ) yy reset y Y (νxy ) xx x, X (νyx ) y clo x νxy xx, X (νyx ) y Y (νxy ) xx, X (νyx ) yy clo y Y (νxy ) xx,νyx y φ(νxy,y (νxy )) xx, φ(x (νyx ),νyx ) y µ Y (νxy ) xx, X (νyx ) y reset x Y (νxy ) x, X (νyx ) yy reset y Y (νxy ) xx, X (νyx ) y clo x νxy x, X (νyx ) y Y (νxy ) x, X (νyx ) yy clo y Y (νxy ) x,νyx y φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ Y (νxy ) x, X (νyx ) y clo y Y (νxy ) x,νyx ε clo x νxy ε,νyx ε 11 / 17

Example: Stirling to circular proofs (IV) Y (νxy ) x, X (νyx ) y reset x Y (νxy ) xx, X (νyx ) y reset y Y (νxy ) xx, X (νyx ) y clo x [νxy xx, X (νyx ) y ] Y (νxy ) xx, X (νyx ) yy clo y [Y (νxy ) xx,νyx y ] φ(νxy,y (νxy )) xx, φ(x (νyx ),νyx ) y µ Y (νxy ) xx, X (νyx ) y reset x Y (νxy ) xx, X (νyx ) y clo x νxy x, X (νyx ) y Y (νxy ) x, X (νyx ) y reset y Y (νxy ) x, X (νyx ) yy clo y [Y (νxy ) x,νyx y ] φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ Y (νxy ) x, X (νyx ) y clo y Y (νxy ) x,νyx ε clo x νxy ε,νyx ε 12 / 17

Example: Stirling to circular proofs (V) A proof in Clo: [νxy xx, X (νyx ) y ] [Y (νxy ) x, X (νyx ) y ] exp Y (νxy ) xx,νyx y φ(νxy,y (νxy )) xx, φ(x (νyx ),νyx ) y µ Y (νxy ) xx, X (νyx ) y clo νxy x, X (νyx ) y [Y (νxy ) x,νyx y ] φ(νxy,y (νxy )) x, φ(x (νyx ),νyx ) y µ Y (νxy ) x, X (νyx ) y clo Y (νxy ) x,νyx ε clo νxy ε,νyx ε 13 / 17

Taking stock We have shown φ valid φ has a tableau Stir φ Clo φ Corollary Circular proofs with ν-closure are complete for the modal µ-calculus Can we get closer to a sequent calculus? no discharge rules no annotations 14 / 17

Eliminating assumptions Let Z = νzψ (z) [Γ 1, Z az 1 ] [Γ k, Z az 1 z k ] (νzψ ) az 1 z i νz(γ i Γ 1 ψ a ) Theorem Γ k,ψ (Z ) az 1 z k clozk Γk, Z az 1 z k 1 Γ i,ψ (Z ) az 1 z i clozi Γ i, Z az 1 z i 1 Γ 1,ψ (Z ) az 1 cloz1 Γ1, Z a Clo Γ Koz +con+ d Γ Γ i,ψ a (νz Γ i Γ 1 ψ a ) d Γ i,ψ a (Γ i νz Γ i Γ 1 ψ a ) + νz Γ i,νz Γ i 1 Γ 1 ψ a (Γ i z)) ind Γ i,νyνz Γ i 1 Γ 1 ψ a (y z)) con Γ i,νz Γ i 1 Γ 1 ψ a Theorem Koz + con + d Γ implies Koz Γ 15 / 17

Summary Clo Stir Valid Koz + d + con Koz 1 We introduce two sound and complete cut-free proof systems; 2 Doing so we provide a new proof of completeness for Koz yielding a procedure for obtaining proofs from tableaux; 3 Koz is complete iff d and con are admissible 16 / 17

Proving almost always implies infinitely often νxy, X (νyx ) Y (νxy ),νyx d Y (νxy ),νyx φ(νxy,y (νxy )), φ(x (νyx ),νyx ) µx + µ y Y (νxy ), X (νyx ) (ν x + d + ind + con) νxy, X (νyx ) Y (νxy ),νyx where φ(νxy,y (νxy )), φ(x (νyx ),νyx ) µx + µ y Y (νxy ), X (νyx ) (ν y + d + ind + con) Y (νxy ),νyx νx νxµyφ,νyµxφ Y (x) = µyφ(x, y) X (y) = µxφ(x, y) Y = X (µyx ) Y X = Y (µxy ) X 17 / 17