Syntactic Complexity of Suffix-Free Languages. Marek Szykuła

Similar documents
CS4445/9544 Analysis of Algorithms II Solution for Assignment 1

FLAT CYCLOTOMIC POLYNOMIALS OF ORDER FOUR AND HIGHER

Problem Set If all directed edges in a network have distinct capacities, then there is a unique maximum flow.

T-Rough Fuzzy Subgroups of Groups

Network Flows: Introduction & Maximum Flow

EECE 301 Signals & Systems Prof. Mark Fowler

Introduction to Congestion Games

Fractional Ornstein-Uhlenbeck Bridge

1 Motivation and Basic Definitions

Average Case Lower Bounds for Monotone Switching Networks

The multisubset sum problem for finite abelian groups

On the Exponential Operator Functions on Time Scales

Flow networks. Flow Networks. A flow on a network. Flow networks. The maximum-flow problem. Introduction to Algorithms, Lecture 22 December 5, 2001

Generalized Orlicz Spaces and Wasserstein Distances for Convex-Concave Scale Functions

Main Reference: Sections in CLRS.

6.8 Laplace Transform: General Formulas

Algorithmic Discrete Mathematics 6. Exercise Sheet

NECESSARY AND SUFFICIENT CONDITIONS FOR LATENT SEPARABILITY

FUZZY n-inner PRODUCT SPACE

18 Extensions of Maximum Flow

Introduction to SLE Lecture Notes

FIXED POINTS AND STABILITY IN NEUTRAL DIFFERENTIAL EQUATIONS WITH VARIABLE DELAYS

CHAPTER 7. Definition and Properties. of Laplace Transforms

Algorithms and Data Structures 2011/12 Week 9 Solutions (Tues 15th - Fri 18th Nov)

Lower and Upper Approximation of Fuzzy Ideals in a Semiring

Graphs III - Network Flow

Randomized Perfect Bipartite Matching

Note on Matuzsewska-Orlich indices and Zygmund inequalities

CHAPTER 7: SECOND-ORDER CIRCUITS

CS 473G Lecture 15: Max-Flow Algorithms and Applications Fall 2005

Network Flow. Data Structures and Algorithms Andrei Bulatov

u(t) Figure 1. Open loop control system

arxiv: v2 [cs.cg] 30 Apr 2015

Let. x y. denote a bivariate time series with zero mean.

, the. L and the L. x x. max. i n. It is easy to show that these two norms satisfy the following relation: x x n x = (17.3) max

Lecture 21. The Lovasz splitting-off lemma Topics in Combinatorial Optimization April 29th, 2004

Spider Diagrams: a diagrammatic reasoning system

CSC 364S Notes University of Toronto, Spring, The networks we will consider are directed graphs, where each edge has associated with it

of Manchester The University COMP14112 Hidden Markov Models

Spider Diagrams: a diagrammatic reasoning system

Linear Algebra Primer

2. VECTORS. R Vectors are denoted by bold-face characters such as R, V, etc. The magnitude of a vector, such as R, is denoted as R, R, V

Challenge Problems. DIS 203 and 210. March 6, (e 2) k. k(k + 2). k=1. f(x) = k(k + 2) = 1 x k

ARTIFICIAL INTELLIGENCE. Markov decision processes

4/12/12. Applications of the Maxflow Problem 7.5 Bipartite Matching. Bipartite Matching. Bipartite Matching. Bipartite matching: the flow network

A Logic of Orthogonality

arxiv: v1 [cs.cg] 21 Mar 2013

The Residual Graph. 12 Augmenting Path Algorithms. Augmenting Path Algorithm. Augmenting Path Algorithm

Admin MAX FLOW APPLICATIONS. Flow graph/networks. Flow constraints 4/30/13. CS lunch today Grading. in-flow = out-flow for every vertex (except s, t)

Group C*-algebras A.K. Recall 1 that if G is a nonempty set, the linear space

CHAPTER 7: UNCERTAINTY

Positive and negative solutions of a boundary value problem for a

The Residual Graph. 11 Augmenting Path Algorithms. Augmenting Path Algorithm. Augmenting Path Algorithm

A LOGIC OF ORTHOGONALITY

Greedy. I Divide and Conquer. I Dynamic Programming. I Network Flows. Network Flow. I Previous topics: design techniques

Bilal Khan Department of Mathematics and Computer Science, John Jay College of Criminal Justice, City University of New York, New York, NY 10019, USA.

1 Adjusted Parameters

More on ODEs by Laplace Transforms October 30, 2017

A Guided Tour in the Topos of Graphs

THE BERNOULLI NUMBERS. t k. = lim. = lim = 1, d t B 1 = lim. 1+e t te t = lim t 0 (e t 1) 2. = lim = 1 2.

SYNTACTIC COMPLEXITY OF R- AND J-TRIVIAL REGULAR LANGUAGES

Longest Common Prefixes

PHYSICS Solving Equations

Research Article On Double Summability of Double Conjugate Fourier Series

Linking Prefixes and Suffixes for Constraints Encoded Using Automata with Accumulators

arxiv: v2 [cs.fl] 6 Mar 2017

Exponential Sawtooth

ANSWERS TO ODD NUMBERED EXERCISES IN CHAPTER

Stability in Distribution for Backward Uncertain Differential Equation

Network Flows UPCOPENCOURSEWARE number 34414

EE202 Circuit Theory II

Chapter 6. Laplace Transforms

Syntactic Complexity of Ideal and Closed Languages

Physics 240: Worksheet 16 Name

arxiv:math/ v2 [math.fa] 30 Jul 2006

ANALYSIS OF SOME SAFETY ASSESSMENT STANDARD ON GROUNDING SYSTEMS

16 Max-Flow Algorithms and Applications

CONTROL SYSTEMS. Chapter 10 : State Space Response

When analyzing an object s motion there are two factors to consider when attempting to bring it to rest. 1. The object s mass 2. The object s velocity

KEY. Math 334 Midterm I Fall 2008 sections 001 and 003 Instructor: Scott Glasgow

Energy Equality and Uniqueness of Weak Solutions to MHD Equations in L (0,T; L n (Ω))

Matching. Slides designed by Kevin Wayne.

23.5. Half-Range Series. Introduction. Prerequisites. Learning Outcomes

5.2 GRAPHICAL VELOCITY ANALYSIS Polygon Method

A NOTE ON PRE-J-GROUP RINGS

Some Ramsey results for the n-cube

ON FRACTIONAL ORNSTEIN-UHLENBECK PROCESSES

An introduction to the (local) martingale problem

Discussion Session 2 Constant Acceleration/Relative Motion Week 03

Lie Derivatives operator vector field flow push back Lie derivative of

13.1 Circuit Elements in the s Domain Circuit Analysis in the s Domain The Transfer Function and Natural Response 13.

Reminder: Flow Networks

To become more mathematically correct, Circuit equations are Algebraic Differential equations. from KVL, KCL from the constitutive relationship

5.1 - Logarithms and Their Properties

Chapter 6. Laplace Transforms

Stationary Distribution. Design and Analysis of Algorithms Andrei Bulatov

23.2. Representing Periodic Functions by Fourier Series. Introduction. Prerequisites. Learning Outcomes

Research Article Existence and Uniqueness of Solutions for a Class of Nonlinear Stochastic Differential Equations

Traversal of a subtree is slow, which affects prefix and range queries.

Stochastic Optimal Control in Infinite Dimensions: Dynamic Programming and HJB Equations

Transcription:

Inroducion Upper Bound on Synacic Complexiy of Suffix-Free Language Univeriy of Wrocław, Poland Join work wih Januz Brzozowki Univeriy of Waerloo, Canada DCFS, 25.06.2015

Abrac Inroducion Sae and ynacic complexiy Suffix-free language We udy ynacic complexiy of he cla of uffix-free language: A language L i uffix-free if w = uv L, where u i non-empy, implie ha he uffix v L. Conribuion Brzozowki, Li, and Ye (TCS 2012) conjecured ha: For n 6, he ynacic complexiy of uffix-free language i (n 1) n 2 + n 2. For n 6, he ynacic complexiy of bifix-free language i (n 1) n 3 +(n 2) n 3 +(n 3)2 n 3. For n 6, he ynacic complexiy of facor-free language i (n 1) n 3 +(n 3)2 n 3 + 1. We prove he fir conjecure (abou uffix-free language).

Lef quoien Inroducion Sae and ynacic complexiy Suffix-free language The (lef) quoien of a regular language L by a word w i w 1 L = {x Σ wx L}. Analogouly, for a ae q of a minimal DFA A = (Q,Σ,δ,q 0,F) recognizing L: L q = {x Σ q x F}, where x i he ranformaion induced by word x. So L q i he e of word aking q o an acceping ae.

Lef quoien Inroducion Sae and ynacic complexiy Suffix-free language The (lef) quoien of a regular language L by a word w i w 1 L = {x Σ wx L}. Analogouly, for a ae q of a minimal DFA A = (Q,Σ,δ,q 0,F) recognizing L: L q = {x Σ q x F}, where x i he ranformaion induced by word x. So L q i he e of word aking q o an acceping ae.

Lef quoien Inroducion Sae and ynacic complexiy Suffix-free language The (lef) quoien of a regular language L by a word w i w 1 L = {x Σ wx L}. Analogouly, for a ae q of a minimal DFA A = (Q,Σ,δ,q 0,F) recognizing L: L q = {x Σ q x F}, where x i he ranformaion induced by word x. So L q i he e of word aking q o an acceping ae.

Sae complexiy Inroducion Sae and ynacic complexiy Suffix-free language Nerode righ congruence on Σ For a regular language L and word x,y Σ : x L y if and only if xv L yv L, for all v Σ Sae complexiy The ae complexiy or quoien complexiy κ(l) of a regular language L i: The number of equivalence clae of L. The number of lef quoien of L. The number of ae in a minimal DFA recognizing L.

Synacic complexiy Inroducion Sae and ynacic complexiy Suffix-free language Myhill congruence For a regular language L and word x,y Σ : x L y if and only if uxv L uyv L for all u,v Σ Synacic complexiy The ynacic complexiy σ(l) of a regular language L i: Σ + / L he number of equivalence clae of L. The ize of he ynacic emigroup of L. The ize of he raniion emigroup of a minimal DFA recognizing L.

Inroducion Sae and ynacic complexiy Suffix-free language Synacic complexiy of a cla of language The ynacic complexiy of a cla of language i: The ize of he large ynacic emigroup of language in ha cla. Expreed a a funcion of he ae complexiie n = κ(l) of he language. In oher word Suppoe we have an n-ae minimal DFA recognizing ome language from he given cla. We ak how many ranformaion (a mo) can be in he raniion emigroup of he DFA.

Inroducion Sae and ynacic complexiy Suffix-free language Synacic complexiy of a cla of language The ynacic complexiy of a cla of language i: The ize of he large ynacic emigroup of language in ha cla. Expreed a a funcion of he ae complexiie n = κ(l) of he language. In oher word Suppoe we have an n-ae minimal DFA recognizing ome language from he given cla. We ak how many ranformaion (a mo) can be in he raniion emigroup of he DFA.

Inroducion Sae and ynacic complexiy Suffix-free language Propoiion n 1 σ(l) n n The bound are igh for n > 1 in he cla of all regular language.

Previou reul Inroducion Sae and ynacic complexiy Suffix-free language Gome, Howie 1992: (parially) monoonic emigroup. Krawez, Lawrence, Shalli 2003: unary and binary alphabe. Holzer, König 2004: unary and binary alphabe. Brzozowki, Ye 2010: ideal and cloed language. Beaudry, Holzer 2011: emigroup of reverible DFA. Brzozowki, Liu 2012: finie, cofinie, definie, revere definie language. Brzozowki, Li, Ye 2012: prefix-, uffix-, bifix-, facor-free language. Iván, Nagy-György 2013: (generalized) definie language. Brzozowki, Li 2013: J -rivial and R-rivial language. Brzozowki, Li, Liu 2013: aperiodic, nearly monoonic emigroup. Brzozowki, Szykuła 2014: aperiodic emigroup. Brzozowki, Szykuła 2014: lef and wo-ided ideal language.

Suffix-free language Inroducion Sae and ynacic complexiy Suffix-free language Suffix-free language A language L i uffix-free if w = uv L, where u i non-empy, implie ha he uffix v L. Baic properie L ha he empy quoien. For w,x Σ +, if w 1 L hen w 1 L (xw) 1 L. For w Σ, he chain of quoien w 1 L,(ww) 1 L,... end in.

Inroducion Sae and ynacic complexiy Suffix-free language Synacic complexiy of uffix-free language Brzozowki, Li, Ye 2012 Here Two maximal emigroup were inroduced: W 5 (n) and W 6 (n). For n 5 he large raniion emigroup i W 5 (n). For n = 6 a large raniion emigroup i W 6 (n). I wa conjecured ha W 6 (n) i alo large for all n 7. The ize of W 6 (n) i (n 1) n 2 +(n 2). We have proved ha W 6 (n) i he unique large raniion emigroup for n 7.

Inroducion Sae and ynacic complexiy Suffix-free language Synacic complexiy of uffix-free language Brzozowki, Li, Ye 2012 Here Two maximal emigroup were inroduced: W 5 (n) and W 6 (n). For n 5 he large raniion emigroup i W 5 (n). For n = 6 a large raniion emigroup i W 6 (n). I wa conjecured ha W 6 (n) i alo large for all n 7. The ize of W 6 (n) i (n 1) n 2 +(n 2). We have proved ha W 6 (n) i he unique large raniion emigroup for n 7.

Inroducion Sae and ynacic complexiy Suffix-free language Properie of DFA of uffix-free language Le D = (Q,Σ,δ,0,F) be a minimal DFA of a uffix-free language. Q = {0,1,...,n 1}. 0 i he ar ae. n 1 i he empy ae. Le T n be he raniion emigroup of D. Baic propery For any ranformaion T n and any ae q Q \{0}, we have 0 q unle 0 = q = n 1.

Inroducion Sae and ynacic complexiy Suffix-free language Properie of DFA of uffix-free language Le D = (Q,Σ,δ,0,F) be a minimal DFA of a uffix-free language. Q = {0,1,...,n 1}. 0 i he ar ae. n 1 i he empy ae. Le T n be he raniion emigroup of D. Baic propery For any ranformaion T n and any ae q Q \{0}, we have 0 q unle 0 = q = n 1.

Colliding pair of ae Inroducion Sae and ynacic complexiy Suffix-free language Definiion Two ae p,q Q are colliding in T n, if here i a ranformaion T n uch ha: 0 = p, r = q for ome ae r Q \{0,n 1}. No ranformaion can map a pair of colliding ae o a ingle ae, excep o n 1. u r q 0 p... n 1

Colliding pair of ae Inroducion Sae and ynacic complexiy Suffix-free language Definiion Two ae p,q Q are colliding in T n, if here i a ranformaion T n uch ha: 0 = p, r = q for ome ae r Q \{0,n 1}. No ranformaion can map a pair of colliding ae o a ingle ae, excep o n 1. u r q 0 p... n 1

Colliding ae Inroducion Sae and ynacic complexiy Suffix-free language Remark In W 5 (n) all pair p,q Q \{0,n 1} are colliding. In W 6 (n) here are no colliding pair.

Inroducion The wine (lower bound) The proof of he upper bound Wine DFA wih he raniion emigroup W 6 (n) generaed by 5 leer. c c, d a, b 0 e 1 b 2 a... a n 2 a, c b, c, d b, d Σ\{e} d, e e e n 1 Σ

Inroducion The wine (lower bound) The proof of he upper bound Le S n = W 6 (n) be he raniion emigroup of he wine The raniion emigroup S n S n conain: All ranformaion ha map 0 o n 1. All ranformaion ha map 0 o a ae in {1,...,n 2}, and all oher ae o n 1. Thee are (n 1) n 2 +(n 2) ranformaion in oal.

Inroducion The wine (lower bound) The proof of he upper bound Idea We aume n 7. S n i he raniion emigroup of he wine. T n i he raniion emigroup of an arbirary DFA of a uffix-free language. We how ha T n S n = (n 1) n 2 +(n 2). I i poible ha T n S n. We conruc an injecive funcion: ϕ: T n S n (Injecive: for every we have ϕ() ϕ( ).)

Inroducion The wine (lower bound) The proof of he upper bound Idea We aume n 7. S n i he raniion emigroup of he wine. T n i he raniion emigroup of an arbirary DFA of a uffix-free language. We how ha T n S n = (n 1) n 2 +(n 2). I i poible ha T n S n. We conruc an injecive funcion: ϕ: T n S n (Injecive: for every we have ϕ() ϕ( ).)

Inroducion The wine (lower bound) The proof of he upper bound Idea We aume n 7. S n i he raniion emigroup of he wine. T n i he raniion emigroup of an arbirary DFA of a uffix-free language. We how ha T n S n = (n 1) n 2 +(n 2). I i poible ha T n S n. We conruc an injecive funcion: ϕ: T n S n (Injecive: for every we have ϕ() ϕ( ).)

Cae 1 Inroducion The wine (lower bound) The proof of he upper bound If S n, hen le ϕ() =. Thi i obviouly injecive, and ϕ() T n. From now, for S n we need o aign a ranformaion ϕ() T n.

Cae 1 Inroducion The wine (lower bound) The proof of he upper bound If S n, hen le ϕ() =. Thi i obviouly injecive, and ϕ() T n. From now, for S n we need o aign a ranformaion ϕ() T n.

Cae 2 Inroducion The wine (lower bound) The proof of he upper bound If S n and ha a cycle, hen we revere he chain p,p,... and map p o a minimal ae ha appear in cycle. : r r... 0 p... p k n 1 : r r... 0 p... p k n 1

Cae 2 Inroducion The wine (lower bound) The proof of he upper bound If S n and ha a cycle, hen we revere he chain p,p,... and map p o a minimal ae ha appear in cycle. : r r... 0 p... p k n 1 : r r... 0 p... p k n 1

Cae 3 Inroducion The wine (lower bound) The proof of he upper bound If S n and ha no cycle, hen we revere he chain and make p a fixed poin. : 0 p... p k n 1 : 0 p... p k n 1

Cae 11 wih p < r Inroducion The wine (lower bound) The proof of he upper bound : q r r 1 r 2 0 p n 1 : q r r 1 r 2 0 p n 1

Cae 11 wih p > r Inroducion The wine (lower bound) The proof of he upper bound : q r r 1 r 2 0 p n 1 : q r r 1 r 2 0 p n 1

The la cae 12 Inroducion The wine (lower bound) The proof of he upper bound : : f r 1 r 2... 0 p n 1 f r 1 r 2... 0 p n 1

We have 12 cae Inroducion The wine (lower bound) The proof of he upper bound Cae 1 S n. Cae 2 ha a cycle. Cae 3 p n 1. Cae 4 There i a fixed poin r Q \{0, n 1} wih in-degree 2. Cae 5 There i r wih in-degree 1 ha i no a fixed poin and r n 1. Cae 6 There i r Q \{0, n 1} wih in-degree 2. Cae 7 There are q 1,q 2 Q \{0, n 1} ha are no fixed poin and aify q 1 n 1 and q 2 n 1. Cae 8 There are wo fixed poin r 1 and r 2 in Q \{0, n 1} wih in-degree 1. Cae 9 There i q Q \{0, n 1} ha i no a fixed poin and aifie q n 1, p < q, and a fixed poin f n 1. Cae 10 There i q Q \{0, n 1} ha i no a fixed poin and aifie q n 1, and a fixed poin f n 1. Cae 11 There i q Q \{0, n 1} ha i no a fixed poin and aifie q n 1. Cae 12 Any oher ranformaion. In each cae we aume ha doe no belong o he previou cae.

Summary Inroducion The wine (lower bound) The proof of he upper bound The 12 cae cover all poibiliie for. Funcion ϕ i injecive and ϕ(t n ) S n, and T n = ϕ(t n ) S n.

Inroducion Uniquene of maximaliy The wine (lower bound) The proof of he upper bound Theorem For n 7, he raniion emigroup S n of he wine i he only one reaching he upper bound. Theorem Five leer are neceary o generae S n. So five leer are needed o reach he upper bound.

Fuure work Inroducion The wine (lower bound) The proof of he upper bound Oher problem The echnique of injecive funcion can be applied o imilar problem. Earlier, we olved he problem of ynacic complexiy of lef ideal (5 ubcae) and wo-ided ideal (8 ubcae) (DLT 2014). The problem of upper bound for ynacic complexiy of bifix- and facor-free language remain open. Thank you!

Fuure work Inroducion The wine (lower bound) The proof of he upper bound Oher problem The echnique of injecive funcion can be applied o imilar problem. Earlier, we olved he problem of ynacic complexiy of lef ideal (5 ubcae) and wo-ided ideal (8 ubcae) (DLT 2014). The problem of upper bound for ynacic complexiy of bifix- and facor-free language remain open. Thank you!