Uniform 2D-Monotone Minimum Spanning Graphs

Similar documents
Planar Upward Drawings

COMPLEXITY OF COUNTING PLANAR TILINGS BY TWO BARS

Paths. Connectivity. Euler and Hamilton Paths. Planar graphs.

An undirected graph G = (V, E) V a set of vertices E a set of unordered edges (v,w) where v, w in V

Outline. 1 Introduction. 2 Min-Cost Spanning Trees. 4 Example

Why the Junction Tree Algorithm? The Junction Tree Algorithm. Clique Potential Representation. Overview. Chris Williams 1.

Algorithmic and NP-Completeness Aspects of a Total Lict Domination Number of a Graph

Constructive Geometric Constraint Solving

Outline. Computer Science 331. Computation of Min-Cost Spanning Trees. Costs of Spanning Trees in Weighted Graphs

Garnir Polynomial and their Properties

Cycles and Simple Cycles. Paths and Simple Paths. Trees. Problem: There is No Completely Standard Terminology!

Graph Isomorphism. Graphs - II. Cayley s Formula. Planar Graphs. Outline. Is K 5 planar? The number of labeled trees on n nodes is n n-2

CSC Design and Analysis of Algorithms. Example: Change-Making Problem

Graphs. Graphs. Graphs: Basic Terminology. Directed Graphs. Dr Papalaskari 1

, each of which is a tree, and whose roots r 1. , respectively, are children of r. Data Structures & File Management

CSE 373: More on graphs; DFS and BFS. Michael Lee Wednesday, Feb 14, 2018

0.1. Exercise 1: the distances between four points in a graph

12/3/12. Outline. Part 10. Graphs. Circuits. Euler paths/circuits. Euler s bridge problem (Bridges of Konigsberg Problem)

5/9/13. Part 10. Graphs. Outline. Circuits. Introduction Terminology Implementing Graphs

The University of Sydney MATH2969/2069. Graph Theory Tutorial 5 (Week 12) Solutions 2008

(2) If we multiplied a row of B by λ, then the value is also multiplied by λ(here lambda could be 0). namely

Exam 1 Solution. CS 542 Advanced Data Structures and Algorithms 2/14/2013

COMP108 Algorithmic Foundations

b. How many ternary words of length 23 with eight 0 s, nine 1 s and six 2 s?

CS 241 Analysis of Algorithms

Math 61 : Discrete Structures Final Exam Instructor: Ciprian Manolescu. You have 180 minutes.

Graphs. CSC 1300 Discrete Structures Villanova University. Villanova CSC Dr Papalaskari

Module graph.py. 1 Introduction. 2 Graph basics. 3 Module graph.py. 3.1 Objects. CS 231 Naomi Nishimura

1 Introduction to Modulo 7 Arithmetic

CS 461, Lecture 17. Today s Outline. Example Run

CSE 373. Graphs 1: Concepts, Depth/Breadth-First Search reading: Weiss Ch. 9. slides created by Marty Stepp

Outline. Circuits. Euler paths/circuits 4/25/12. Part 10. Graphs. Euler s bridge problem (Bridges of Konigsberg Problem)

Trees as operads. Lecture A formalism of trees

Solutions for HW11. Exercise 34. (a) Use the recurrence relation t(g) = t(g e) + t(g/e) to count the number of spanning trees of v 1

NP-Completeness. CS3230 (Algorithm) Traveling Salesperson Problem. What s the Big Deal? Given a Problem. What s the Big Deal? What s the Big Deal?

V={A,B,C,D,E} E={ (A,D),(A,E),(B,D), (B,E),(C,D),(C,E)}

Weighted graphs -- reminder. Data Structures LECTURE 15. Shortest paths algorithms. Example: weighted graph. Two basic properties of shortest paths

V={A,B,C,D,E} E={ (A,D),(A,E),(B,D), (B,E),(C,D),(C,E)}

MAT3707. Tutorial letter 201/1/2017 DISCRETE MATHEMATICS: COMBINATORICS. Semester 1. Department of Mathematical Sciences MAT3707/201/1/2017

A 4-state solution to the Firing Squad Synchronization Problem based on hybrid rule 60 and 102 cellular automata

CS200: Graphs. Graphs. Directed Graphs. Graphs/Networks Around Us. What can this represent? Sometimes we want to represent directionality:

Properties of Hexagonal Tile local and XYZ-local Series

ECE COMBINATIONAL BUILDING BLOCKS - INVEST 13 DECODERS AND ENCODERS

Present state Next state Q + M N

Section 10.4 Connectivity (up to paths and isomorphism, not including)

S i m p l i f y i n g A l g e b r a SIMPLIFYING ALGEBRA.

Similarity Search. The Binary Branch Distance. Nikolaus Augsten.

Polygons POLYGONS.

QUESTIONS BEGIN HERE!

Register Allocation. Register Allocation. Principle Phases. Principle Phases. Example: Build. Spills 11/14/2012

CS September 2018

Announcements. Not graphs. These are Graphs. Applications of Graphs. Graph Definitions. Graphs & Graph Algorithms. A6 released today: Risk

Solutions to Homework 5

Numbering Boundary Nodes

Designing A Concrete Arch Bridge

On Local Transformations in Plane Geometric Graphs Embedded on Small Grids

QUESTIONS BEGIN HERE!

arxiv: v1 [math.co] 15 Dec 2015

CSE 373: AVL trees. Warmup: Warmup. Interlude: Exploring the balance invariant. AVL Trees: Invariants. AVL tree invariants review

Outline. Binary Tree

Steinberg s Conjecture is false

Problem solving by search

Using the Printable Sticker Function. Using the Edit Screen. Computer. Tablet. ScanNCutCanvas

GREEDY TECHNIQUE. Greedy method vs. Dynamic programming method:

MULTIPLE-LEVEL LOGIC OPTIMIZATION II

Research Article On the Genus of the Zero-Divisor Graph of Z n

Computational Biology, Phylogenetic Trees. Consensus methods

CS61B Lecture #33. Administrivia: Autograder will run this evening. Today s Readings: Graph Structures: DSIJ, Chapter 12

Chapter 18. Minimum Spanning Trees Minimum Spanning Trees. a d. a d. a d. f c

Nefertiti. Echoes of. Regal components evoke visions of the past MULTIPLE STITCHES. designed by Helena Tang-Lim

Witness-Bar Visibility Graphs

Analysis for Balloon Modeling Structure based on Graph Theory

12. Traffic engineering

Discovering Pairwise Compatibility Graphs

FSA. CmSc 365 Theory of Computation. Finite State Automata and Regular Expressions (Chapter 2, Section 2.3) ALPHABET operations: U, concatenation, *

Greedy Algorithms, Activity Selection, Minimum Spanning Trees Scribes: Logan Short (2015), Virginia Date: May 18, 2016

# 1 ' 10 ' 100. Decimal point = 4 hundred. = 6 tens (or sixty) = 5 ones (or five) = 2 tenths. = 7 hundredths.

Announcements. These are Graphs. This is not a Graph. Graph Definitions. Applications of Graphs. Graphs & Graph Algorithms

Construction 11: Book I, Proposition 42

EE1000 Project 4 Digital Volt Meter

Graph Contraction and Connectivity

Register Allocation. How to assign variables to finitely many registers? What to do when it can t be done? How to do so efficiently?

A comparison of routing sets for robust network design

RAM Model. I/O Model. Real Machine Example: Nehalem : Algorithms in the Real World 4/9/13

More Foundations. Undirected Graphs. Degree. A Theorem. Graphs, Products, & Relations

Straight-line Grid Drawings of 3-Connected 1-Planar Graphs

A 43k Kernel for Planar Dominating Set using Computer-Aided Reduction Rule Discovery

CS553 Lecture Register Allocation 1

DUET WITH DIAMONDS COLOR SHIFTING BRACELET By Leslie Rogalski

5/7/13. Part 10. Graphs. Theorem Theorem Graphs Describing Precedence. Outline. Theorem 10-1: The Handshaking Theorem

learning objectives learn what graphs are in mathematical terms learn how to represent graphs in computers learn about typical graph algorithms

A Simple Code Generator. Code generation Algorithm. Register and Address Descriptors. Example 3/31/2008. Code Generation

Page 1. Question 19.1b Electric Charge II Question 19.2a Conductors I. ConcepTest Clicker Questions Chapter 19. Physics, 4 th Edition James S.

Seven-Segment Display Driver

Can transitive orientation make sandwich problems easier?

Minimum Spanning Trees

10/30/12. Today. CS/ENGRD 2110 Object- Oriented Programming and Data Structures Fall 2012 Doug James. DFS algorithm. Reachability Algorithms

IEEE TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS, VOL. TK, NO. TK, MONTHTK YEARTK 1. Hamiltonian Walks of Phylogenetic Treespaces

XML and Databases. Outline. Recall: Top-Down Evaluation of Simple Paths. Recall: Top-Down Evaluation of Simple Paths. Sebastian Maneth NICTA and UNSW

This chapter covers special properties of planar graphs.

Transcription:

CCCG 2018, Winnipg, Cn, August 8 10, 2018 Uniorm 2D-Monoton Minimum Spnning Grphs Konstntinos Mstks Astrt A gomtri grph G is xy monoton i h pir o vrtis o G is onnt y xy monoton pth. W stuy th prolm o prouing th xy monoton spnning gomtri grph o point st P tht (i) hs th minimum ost, whr th ost o gomtri grph is th sum o th Eulin lngths o its gs, n (ii) hs th lst numr o gs, in th ss tht th Crtsin Systm xy is spii or rly slt. Builing upon prvious rsults, w sily otin tht th two solutions oini whn th Crtsin Systm is spii n r oth qul to th rtngl o inlun grph o P. Th rtngl o inlun grph o P is th gomtri grph with vrtx st P suh tht two points p, q P r jnt i n only i th rtngl with ornrs p n q os not inlu ny othr point o P. Whn th Crtsin Systm n rly hosn, w not tht th two solutions o not nssrily oini, howvr w show tht thy n oth otin in O( P 3 ) tim. W lso giv simpl 2 pproximtion lgorithm or th prolm o omputing th spnning gomtri grph o k root point st P, in whih h root is onnt to ll th othr points (inluing th othr roots) o P y y monoton pths, tht hs th minimum ost. 1 Introution A squn o points in th Eulin pln q 0, q 1,..., q t is ll y monoton i th squn o thir y oorints, i.. y(q 0 ), y(q 1 ),..., y(q t ), is ithr rsing or inrsing, with y(p) noting th y oorint o th point p. A gomtri pth Q = (q 0, q 1,..., q t ) is ll y monoton i th squn o its vrtis, i.. th squn q 0, q 1,..., q t, is y monoton. I Q is y monoton or som xis y thn Q is ll monoton. Lt G = (P, E) gomtri grph. I h p, q P r onnt y y monoton pth thn G is ll y monoton. I G is y monoton or som xis y thn G is ll uniorm monoton (ollowing th trminology o [22]). Uniorm monoton grphs wr ll This rsrh ws innilly support y th Spil Aount or Rsrh Grnts o th Ntionl Thnil Univrsity o Athns. Shool o Appli Mthmtil n Physil Sins, Ntionl Thnil Univrsity o Athns, Athns, Gr, kmst@mth.ntu.gr 1 monoton grphs y Anglini [3]. I h p, q P r onnt y monoton pth, whr th irtion o monotoniity might ir or irnt pirs o vrtis, thn G is ll monoton. Monoton grphs wr introu y Anglini t l. [4]. Drwing n (strt) grph s monoton (gomtri) grph hs n topi o rsrh [3, 4, 5, 13, 24]. Th Monoton Minimum Spnning Grph prolm, i.. th prolm o onstruting th monoton spnning gomtri grph o givn point st tht hs th minimum ost, whr th ost o gomtri grph is th sum o th Eulin lngths o its gs, ws rntly introu (ut not solv) in [22] n it rmins n opn prolm whthr it is NP-hr. Sin th mor gnrl (without th rquirmnt o monotoniity) Eulin Minimum Spnning Tr prolm n solv in Θ( P log P ) tim [27], this onstituts grt irntition tht is inu y th ition o th proprty o monotoniity. A point st P is k root i thr xist k points r 1, r 2,..., r k P istinguish rom th othr points o P whih r ll th roots o P. A gomtri grph G = (P, E) is ll k root i P is k root n its roots r th roots o P. A k root gomtri grph G is k root y monoton i h root r P n h point p P \ {r} r onnt y y monoton pths. Similrly, G is k root uniorm monoton (ollowing th trminology o [22]) i it is k root y monoton or som xis y. For simpliity, w my lso not point sts or gomtri grphs tht r 1 root simply s root. A polygon tht is 2 root y monoton, in whih its roots r its lowst n highst vrtis, n tringult in linr tim [11]. L n Prprt [16] prpross suivision S o th pln suh tht th rgion in whih qury point longs n oun quikly, y (i) xtning th gomtri grph ouning S to 2 root y monoton plnr gomtri grph in whih th roots r th highst n lowst vrtis o S, n (ii) onstruting st o pproprit y monoton pths rom th lowst to th highst vrtx o S. Aitionlly, L n Prprt [16] not tht 2 root plnr gomtri grph, whr ll vrtis hv irnt y oorints, in whih th roots r th highst n lowst vrtis o th grph is 2 root y monoton i n only i h non-root vrtx hs oth nighor ov it n nighor low it. Furthrmor, root gomtri grph G = (P, E), whr ll vrtis hv irnt y oorints, with

30 th Cnin Conrn on Computtionl Gomtry, 2018 (singl) root r tht is not th highst or lowst point o P is root y monoton i n only i h non-root vrtx p hs nighor q suh tht y(q) is twn y(r) (inlusiv) n y(p) [22]. Aitionlly, root uniorm monoton grphs n iintly rogniz [22]. Th k root y monoton (uniorm monoton) minimum spnning grph (ollowing th trminology o [22]) o k root point st P is th k root y monoton (uniorm monoton) spnning grph o P tht hs th minimum ost. Th root y monoton (uniorm monoton) minimum spnning grph 1 o root point st P n prou in O( P log 2 P ) (rsp., O( P 2 log P )) tim [22]. Th prolm o rwing root tr s root y monoton minimum spnning grph is stui in [20]. Th ( P root) y monoton minimum spnning grph o point st P is th gomtri pth tht trvrss ll th points o P y moving north, rom th lowst point to th highst point o P [22]. Rgring th prolm o prouing th k root y monoton minimum spnning grph o k root point st P, with 1 < k < P, it is n opn prolm, pos in [22], whthr it is NP-hr. Th rstrit thrs tr prolm ws introu in [12] n is rlt to th root y-monoton minimum spnning grph prolm onstrin to root point sts P in whih th y oorint o th root is zro n th y oorints o th othr points o P r ll ngtiv (or ll positiv). Th input o th rstrit thrs tr prolm is omplt grph with root whr h g hs ost n h vrtx hs vlu n th gol is to output th spnning tr in whih th pth rom th root to h vrtx rss in vlu tht hs th minimum ost. Th rstrit thrs tr prolm is grily solvl [12, Corollry 2.6]. A gomtri pth Q = (q 0, q 1,..., q t ) is xy monoton i th squn o its vrtis is oth x monoton, i.. th squn x(q 0 ), x(q 1 ),..., x(q t ), is monoton, n y monoton. Q is 2D-monoton (ollowing th trminology o [22]) i it is x y monoton or som orthogonl xs x, y. A gomtri grph G = (P, E) is 2D-monoton (ollowing th trminology o [22]) i h pir o points o P is onnt y 2D-monoton pth. 2D-monoton pths/grphs wr ll ngl-monoton pths/grphs y Bonihon t l. [8]. Bonihon t l. [8] show tht iing i gomtri grph G = (P, E) is 2D-monoton n on in O( P E 2 ) tim. Tringultions with no otus intrnl ngls r 2D-monoton grphs [10, 19]. Thr xist point sts or whih ny 2D-monoton spnning grph is not plnr [8]. Th prolm o onstruting 2D-monoton grphs with symptotilly lss thn qurti gs ws stui y Luiw n Monl [18]. It is n opn prolm, pos in [22], whthr th 2Dmonoton spnning grph o point st P tht hs th 1 In [22] it is shown tht it is tully tr. minimum ost n iintly omput. Th (root) xy monoton n (root) uniorm 2Dmonoton (using th trminology o [22]) grphs r in similr to th (root) y monoton n (root) uniorm monoton grphs. Diing i root gomtri grph G = (P, E) is root xy monoton (uniorm 2D-monoton) n on in O( E ) (rsp., O( E log P )) tim [22]. Aitionlly, th root xy monoton (uniorm 2D-monoton) spnning grph o root point st P tht hs th minimum ost 2 n omput in O( P log 3 P ) (rsp., O( P 2 log P )) tim [22]. W ous on th proution o th xymonoton minimum spnning grph (xy MMSG) o point st P, i.. th xy-monoton spnning grph o P tht hs th minimum ost, n th proution o th uniorm 2D monoton minimum spnning grph (2D UMMSG) o point st P, i.. th uniorm 2D monoton spnning grph o P tht hs th minimum ost. W lso stuy th orrsponing prolms rgring th proution o th spnning grphs with th lst numr o gs, i.. th proution o th xy-monoton spnning grph with th lst numr o gs n th proution o th uniorm 2D monoton spnning grph with th lst numr o gs. A urv C is inrsing-hor [15, 26] i or h p 1, p 2, p 3, p 4 trvrs in this orr long it, th lngth o th lin sgmnt p 1 p 4 is grtr thn or qul to th lngth o p 2 p 3. Almri t l. [1] introu inrsinghor grphs whih r th gomtri grphs or whih h two vrtis r onnt y n inrsing-hor pth. Inrsing-hor grphs r wily stui [1, 6, 10, 21, 23]. Th prolm o prouing inrsing-hor spnning grphs (whr Stinr points my ) o point st P ws stui in [1, 10, 21]. Th pproh mploy in [1, 10, 21], ws to onnt th points o P y 2D-monoton pths sin s not y Almri t l. [1] 2D-monoton pths r lso inrsing-hor pths. Lt P point st n lt p, q P thn p n q r rtngulrly visil i th rtngl with ornrs p n q os not inlu ny othr point o P. Furthrmor, th rtngl o inlun grph o P is th gomtri grph spnning P suh tht pq is n g o th grph i n only i p n q r rtngulrly visil. Alon t l. [2] not rtngulrly visil points s sprt points n th rtngl o inlun grph s th sprtion grph. Computing th rtngl o inlun grph G = (P, E) o P n on in O( P log P + E ) tim [25]. Thr xist point sts P or whih th numr o gs o thir rtngl o inlun grph is Ω( P 2 ) [2]. Th rtngl o inlun grph os not rmin th sm i th Crtsin Systm is rott [14, Proposition 3]. 2 In [22] it is shown tht it is tully tr, not s th root xy monoton (uniorm 2D-monoton) minimum spnning tr in [22] n rvit s th root xy MMST (rsp., root 2D UMMST ) in [22].

CCCG 2018, Winnipg, Cn, August 8 10, 2018 Drwing n strt grph s rtngl o inlun grph hs n stui [17]. Our Contriution. Builing upon prvious rsults, w sily otin tht givn point st P th xy MMSG o P is qul to th xy monoton spnning grph o P tht hs th lst numr o gs n r oth qul to th rtngl o inlun grph o P. W not tht givn point st P th 2D UMMSG o P os not nssrily oini with th uniorm 2D-monoton spnning grph o P tht hs th lst numr o gs. W lso show tht oth th 2D UMMSG o P n th uniorm 2D-monoton spnning grph o P tht hs th lst numr o gs n prou in O( P 3 ) tim. Aitionlly, w giv simpl 2 pproximtion lgorithm or th prolm o prouing th k root y monoton minimum spnning grph o k root point st. 2 Prliminris 2.1 xy Monoton Minimum Spnning Grphs Anglini [3] not th ollowing Ft rgring y monoton grphs. Ft 1 (Anglini [3]) Lt G = (P, E) y monoton grph whr no two points o P hv th sm y oorint n lt p, q P suh tht or h r P \ {p, q} th squn p, r, q is not y monoton. Thn, p n q r jnt in G. Ft 1 is sily xtn in th ontxt o xy monoton grphs. Mor spiilly, lt G = (P, E) xy-monoton grph n p, q P suh tht or h r P \ {p, q} th squn o points p, r, q is not xy monoton, thn p n q r jnt in G. Alon t l. [2] not tht th points p, q o point st P r rtngulrly visil i n only i or h r P \{p, q} th squn o points p, r, q is not xy monoton. Hn, th rtngl o inlun grph o P is sugrph o G. Liott t l. [17, Lmm 2.1] show tht th rtngl o inlun grph o point st is xy monoton grph 3. From th prvious two sntns, rgring th rtngl o inlun grph, w otin th ollowing Corollry. Corollry 1 Lt P point st. Th xy MMSG o P n th xy monoton spnning grph o P tht hs th lst numr o gs oini n thy r oth qul to th rtngl o inlun grph o P. 3 Thnilly spking, Liott t l. [17] show tht th rtngl o inlun grph o point st is grph suh tht h two vrtis r onnt y pth lying insi th rtngl in y ths vrtis ut upon rul ring th pth tht is otin in thir proo is xy monoton. W rll tht th rtngl o inlun grph G = (P, E) o P n prou in O( P log P + E ) tim [25] whih is optiml [25] n tht thr xist point sts P or whih th rtngl o inlun grph hs siz Ω( P 2 ) [2] s wll s point sts or whih it hs linr siz [2]. 2.2 Root Uniorm 2D-Monoton Grphs Mstks n Symvonis [22] stui th prolm o rognizing root uniorm 2D-monoton grphs. Thy initilly not th ollowing Ft. Ft 2 (Osrvtion 8 in [22]) Lt G gomtri grph G = (P, E) with root r. I on rotts Crtsin Systm x y, thn G my om root x y monoton whil prviously it ws not, or vi vrs, only whn th y xis oms (or lvs th position whr it prviously ws) prlll or orthogonl to 1. lin pssing through r n point p P \ {r}. 2. n g pq E, whr p, q r. Bs on Ft 2, Mstks n Symvonis [22] gv rottionl swp lgorithm not s th root uniorm 2D-monoton rognition lgorithm in [22]. Ft 3 ([22]) Th root uniorm 2D-monoton rognition lgorithm i) omputs, in O( E log P ) tim, st o suiint Crtsin Systms, o siz O( E ), whih r ssoit with (1) lins pssing through r n point p P \ {r} n (2) gs pq E, whr p, q r. ii) tsts, in O( E ) totl tim 4, i G is root x y monoton or som Crtsin Systm x y in th prviously omput st o suiint Crtsin Systms. Ft 4 (Thorm 1 in [22]) Lt P root point st thn th root y monoton minimum spnning grph o P n otin in O( P log 2 P ) tim. 3 Th 2D-UMMSG Prolm W now l with th onstrution o th 2D UMMSG n th uniorm 2D monoton spnning grph with 4 Thnilly spking in [22] it is shown tht th rmining stps, i.. th stps tr th omputtion o th suiint Crtsin Systms, o th root uniorm 2D-monoton rognition lgorithm tk O( E log P ) totl tim. Intrnlly in th root uniorm 2D-monoton rognition lgorithm givn in [22], or h p P \ {r} it is stor th st o jnt points to p tht r in th rtngl w.r.t. th Crtsin Systm x y with ornrs p n r, whih is not s A(p, x, y ) in [22]. Furthrmor, it is stor th st o points p P \ {r} or whih A(p, x, y ) > 0 whih is not s B(x, y ) in [22]. Howvr, only th rinlitis o ths sts r nssry [22, Lmm 9], hn i inst o th sts A(p, x, y ), p P n B(x, y ) thir rinlitis r stor, th rmining stps o th root uniorm 2D-monoton rognition lgorithm tk O( E ) totl tim.

30 th Cnin Conrn on Computtionl Gomtry, 2018 th lst numr o gs. W initilly show tht th 2D UMMSG o point st P n otin in O( P 3 ) tim. For this, w mploy rottionl swp thniqu. Our pproh rgring th onstrution o th 2D UMMSG is similr to th pproh mploy or th lultion o th root uniorm 2D monoton spnning grph tht hs th minimum ost in [22]. W ssum tht no thr points o P r ollinr n no two lin sgmnts pq n p q, p, p, q, q P, r prlll or orthogonl. Lt P point st n p point o P. Lt RV (p, x, y ) not th sust o points o P tht r rtngulrly visil rom p w.r.t. th Crtsin Systm x y. S or xmpl, Figur 1(). Proposition 2 I w rott Crtsin Systm x y ountrlokwis, thn th x y MMSG o P hngs only whn y rhs or movs wy rom lin prpniulr or prlll to lin pssing through two points o P. Proo. I w rott th Crtsin Systm x y ountrlokwis thn th RV (p, x, y ) or point p P hngs only whn y rhs or movs wy rom lin prpniulr or prlll to lin pssing through two points o P ;.g. s Figur 1. From th prvious n sin th RV (p, x, y ), p P, quls to th st o jnt vrtis o p in th x y MMSG o P (Corollry 1), w otin th Proposition. y p () g h i y j x p () g h i j x y p () Figur 1: In () RV (p, x, y ) = {,,,, g, h, i}. In () th y oms prlll to th n now is not rtngulrly visil rom p. Finlly, in () th y hs lt th position whr it prviously ws orthogonl to th n now oms rtngulrly visil rom p. Lt S = {s [0, π 2 ) : lin o slop s is prpniulr or prlll to lin pssing through two points o P }. g h i j x Lt S = {s 1, s 2,..., s l } with l = ( ) P 2 suh tht 0 s1 < s 2 <... < s l < π 2. W now in th st S suiint s to qul to {s 1, 1+s 2 s 2, s 2, 2+s 3 s 2,..., s l, l + π 2 2 }. Lt x 1 y 1, x 2 y 2,..., x 2l y 2l th Crtsin Systms in whih th vrtil xis hs slop in S suiint, orr w.r.t. th slop o thir vrtil xis. Thorm 3 Th uniorm 2D monoton minimum spnning grph o point st P n omput in O( P 3 ) tim. Proo. From Proposition 2 n th prvious initions w otin th ollowing Proposition. Proposition 4 Th uniorm 2D monoton minimum spnning grph o P is on o th x y MMSG o P ovr ll Crtsin Systms x y with y o slop in S suiint. W now giv O( P 3 ) tim rottionl swp lgorithm. Th lgorithm initilly omputs th x 1 y 1 MMSG o P n thn it otins h x i+1 y i+1 MMSG o P rom th x i y i MMSG o P. Throughout th prour th Crtsin Systm x opt y opt in whih th lgorithm nountr th minimum ost solution so r is stor. In its lst stp, th lgorithm romputs th x opt y opt MMSG o P, whih sin it is qul to th rtngl o inlun grph G = (P, E) w.r.t. th Crtsin Systm x opt y opt (Corollry 1) it n omput in O( P log P + E ) tim [25]. Th ruil proposition (whih w show ltr) tht mks th tim omplxity o th lgorithm qul to O( P 3 ) is tht h trnsition rom th x i y i MMSG o P to th x i+1 y i+1 MMSG o P tks O( P ) tim. For h two points p, q o P lt I(p, q, x i, y i ) th numr o points o P \ {p, q} tht r inlu in th rtngl w.r.t. th Crtsin Systm x i y i with opposit vrtis p n q. Thn, RV (q, x i, y i ) n quivlntly in using th quntitis I(p, q, x i, y i ), p P \ {q}, s ollows: p RV (q, x i, y i ) i I(p, q, x i, y i ) = 0. W stor th RV (q, x i, y i ), q P, i = 1,2,..., 2l in th t strutur rv(q) whih is implmnt s n rry o P oolns. W lso stor th I(p, q, x i, y i ), p, q P, i = 1,2,..., 2l in th vril i(p, q). Computing th Crtsin Systms x i y i, i = 1, 2,..., 2l n on in O( P 2 log P ) tim. Aompni with h Crtsin Systm x i y i is th pir o points (p i, q i ) suh tht p i q i is ithr prlll or prpniulr to th y i xis or th y i 1 xis. Ihino n Sklnsky [14] not tht mploying rng tr [7, 9] tht ontins th points o P on n lult i) th rtngl o inlun grph o P, n ii) th I(p, q, x, y), p, q, P, or Crtsin Systm xy. Applying th prviously mntion pproh, not y Ihino n Sklnsky [14], r otin i) th rtngl o inlun grph o P w.r.t. th Crtsin Systm x 1 y 1 (whih y Corollry 1 quls to th x 1 y 1 MMSG o P ), n ii) th I(p, q, x 1, y 1 ), p, q, P.

CCCG 2018, Winnipg, Cn, August 8 10, 2018 W now show tht w n upt ll th rv(p), p P, suh tht rom qul to RV (p, x i 1, y i 1 ), p P, thy om qul to RV (p, x i, y i ), p P, in O( P ) totl tim. For h p P \ {p i, q i } th upt o rv(p) tks O(1) tim. This is tru, sin only th points p i n q i hv to tst or inlusion to or rmovl rom rv(p). Mor spiilly, w hv to tst i or on o thm, sy p i, th rtngl with ornrs p n p i ontins (or it os not ontin) q i w.r.t. th Crtsin Systm x i y i whil it i not ontin (or it ontin) it w.r.t. x i 1 y i 1. I this is tru, thn th i(p i, p) hngs n p i hs to tst or mmrship in rv(p) n inlu to or rmov rom rv(p). Rgring rv(p i ), th upt tks O( P ) tim, sin or h othr point q P \ {p i, q i } w hv to tst i th rtngl with ornrs q n p i ontins (or it os not ontin) q i w.r.t. th Crtsin Systm x i y i whil it i not ontin it (or it ontin it) w.r.t. th x i 1 y i 1 n i so upt oth th i(q, p i ) n th xistn o q in rv(p i ) i nssry. Similrly, rv(q i ) n upt in O( P ) tim. W not tht th prour o otining th 2D UMMSG n trivilly moii suh tht th uniorm 2D-monoton spnning grph o point st P with th lst numr o gs n otin in O( P 3 ) tim. Sin or n ritrry Crtsin Systm x y th x y MMSG o P is qul to th x y monoton spnning grph o P with th lst numr o gs (Corollry 1), th only moiition whih is nssry is tht in th trnsition rom th Crtsin Systm x i y i to th Crtsin Systm x i+1 y i+1 w hk i th x i+1 y i+1 monoton spnning grph o P with th lst numr o gs hs th lst numr o gs mong ll th prou solutions so r. In Figur 2 is givn point st P or whih th 2D UMMSG o P is irnt rom th uniorm 2D monoton spnning grph o P with th lst numr o gs. In Figur 3 w giv point st P or whih th (nonuniorm) 2D monoton spnning grph o P with th lst numr o gs os not oini to th (nonuniorm) 2D monoton spnning grph o P tht hs th minimum ost. Rgring rognizing uniorm 2D-monoton grphs, w not tht th O( E log P ) tim rottionl swp lgorithm givn in [22], whih is i gomtri grph G = (P, E) with spii vrtx r s root is root uniorm 2D monoton, n sily xtn into O( P 2 log P + P E ) tim rottionl swp lgorithm tht is i G is uniorm 2D monoton. Mor spiilly, in orr to i i G is uniorm 2D monoton, th P root gomtri grphs (p 1, G), (p 2, G),..., (p P, G) whr (p i, G) is th gomtri grph G with root p i n {p 1, p 2,..., p P } is th vrtx st o G, r onsir. A Crtsin Systm x y is rott ountrlokwis. From Ft 2, it ol () () Figur 2: Th points, n orm right ngl. Aitionlly, th points, n orm right ngl. Th slop o is smllr thn th slop o. Th uniorm 2D monoton spnning grph with th lst numr o gs is otin whn th y xis oms prpniulr to th n is shown in (). On th othr hn th 2D UMMSG is otin whn th y xis oms prpniulr to th n is shown in (). () () Figur 3: Th slop o is π 4 whil th slop o is 3π 4. In () is pit th 2D monoton spnning grph o P with th lst numr o gs. In () is illustrt th 2D monoton spnning grph o P tht hs th minimum ost. lows tht on o ths P root gomtri grphs oms root x y monoton whil prviously it ws not, or vi vrs, only whn th y xis oms (or lvs th position whr it ws prviously) prlll or orthogonl to lin pssing through two points o P. Hn, O( P 2 ) Crtsin Systms n to onsir, whih n omput in O( P 2 log P ) tim. Whn th y oms (or lvs th position tht it prviously ws) prlll or prpniulr to lin pssing through th points p, q P thn y Ft 2 th sttus, i.. ing root x y monoton, o th root gomtri grphs (p, G) n (q, G) my hng. Hn, th stps o th root uniorm 2D-monoton rognition lgorithm givn in [22] or hnling th vnt ssoit with th urrnt Crtsin Systm x y rgring th root gomtri grphs (p, G) n (q, G), r ppli. Furthrmor, i pq E thn y Ft 2 it ollows tht th sttus, i.. ing root x y monoton, o h (r, G), r P \ {p, q}, my lso hng. Hn, or h (r, G), r P \ {p, q}, th stps o th root uniorm 2D-monoton rognition lgorithm givn in [22] or hnling th vnt ssoit with th urrnt Crtsin Systm x y r ppli. Sin, th rmining stps, i.. tr th lultion o th suiint xs, o th root uniorm 2D-monoton rognition lgorithm, givn in [22], rgring ny o ths P root

30 th Cnin Conrn on Computtionl Gomtry, 2018 gomtri grphs tk O( E ) tim (Ft 3), pplying th rmining stps rgring ll ths P root gomtri grphs, tks O( P E ) totl tim. 4 A 2 Approximtion Algorithm or th k Root y Monoton Minimum Spnning Grph Prolm W now stuy th prolm o prouing th k root y monoton minimum spnning grph o k root point st P, whr 1 < k < P. W ssum tht no two points hv th sm y oorint. Lt P point st n, R thn P y> is th sust o points o P whos y oorint is grtr thn. Similrly r in P y, P y< n P y. P <y< is th sust o points o P whos y oorint is twn n. Similrly r in P <y, P y< n P y. In [22, Lmm 1] it is not tht th root y monoton minimum spnning grph o root point st P with root r is th union o th root y monoton minimum spnning grphs o (i) P y y(r) n (ii) P y y(r). Th prvious Ft is xtn to th ollowing Lmm. Lmm 5 Lt P k root point st, with 1 < k < P, whr r 1, r 2,..., r k r th roots o P suh tht y(r 1 ) < y(r 2 ) <... < y(r k ). Th k root y monoton minimum spnning grph o P is th union o 1. th root y monoton minimum spnning grph o P y y(r1). 2. th root y monoton minimum spnning grph o P y y(rk ). 3. th 2 root y monoton minimum spnning grph o P y(ri) y y(r i+1), 1 i k 1. Thorm 6 Givn k root point st P, with 1 < k < P, w n otin in O( P log 2 P ) tim k root y monoton spnning grph o P with ost t most twi th ost o th k root y monoton minimum spnning grph o P. Proo. For 2 root point st P with roots r 1 n r 2 tht r th lowst n highst points o th point st, rsptivly, w prov th ollowing Lmm. Lmm 7 Givn 2 root point st P with roots r 1 n r 2 tht r th lowst n highst points o th point st, rsptivly, w n otin in O( P log 2 P ) tim 2 root y monoton spnning grph o P with ost t most twi th ost o th 2 root y monoton minimum spnning grph o P. Proo. Initilly, w mploy Ft 4 to P onsiring it to hv only th root r 1 n otin th gomtri grph G 1. Thn, w mploy Ft 4 to P onsiring it to hv only th root r 2, otining G 2. In th inl stp w rturn th union o G 1 n G 2. G 1 G 2 is 2 root y monoton sin G 1 (G 2 ) is root y monoton with root r 1 (rsp., r 2 ). W now show tht G 1 G 2 hs ost t most twi th ost o th 2 root y monoton minimum spnning grph G opt o P. Sin, in G opt ll th points p r onnt with r 1 (r 2 ) y y monoton pths it ollows tht its ost is grtr thn or qul to th ost o G 1 (rsp., G 2 ). Hn, th ost o G 1 G 2 whih is lss thn or qul to th sum o th osts o G 1 n G 2 is t most twi th ost o G opt. From Lmm 5, Ft 4 n Lmm 7 w otin th Thorm. A 2 root plnr gomtri grph G = (P, E) with roots r 1, r 2 s.t. y(r 1 ) < y(p) < y(r 2 ), p P \ {r 1, r 2 }, is 2 root y monoton i n only i or h p P \ {r 1, r 2 } thr xist q 1, q 2 Aj(p) with y(q 1 ) < y(p) < y(q 2 ) [16]. Furthrmor, root gomtri grph G = (P, E) with (singl) root r tht is not th highst or lowst point o P is root y monoton i n only i or h p P \ {r} thr xists q Aj(p) suh tht y(q) is twn y(r) (inlusiv) n y(p) [22]. W xtn th prvious two Propositions to th ollowing quivlnt hrtriztion o k root y monoton grphs whr th lttr implis n iint rognition lgorithm or k root y monoton grphs. Proposition 8 Lt G = (P, E) k root gomtri grph, whr 1 < k < P, with roots r 1, r 2,..., r k suh tht y(r 1 ) < y(r 2 ) <... < y(r k ). G is k root y monoton i n only i 1. or h p P y<y(r1) thr xists q Aj(p) s.t. y(q) (y(p), y(r 1 )]. 2. or h p P y>y(rk ) thr xists q Aj(p) s.t. y(q) [y(r k ), y(p)). 3. or h p P y(ri)<y<y(r i+1) thr xist q 1, q 2 Aj(p) s.t. y(q 1 ) [y(r i ), y(p)) n y(q 2 ) (y(p), y(r i+1 )], i = 1, 2,..., k 1. 4. thr xists q Aj(r 1 ) s.t. y(q) (y(r 1 ), y(r 2 )]. 5. thr xists q Aj(r k ) s.t. y(q) [y(r k 1 ), y(r k )). 6. thr xist q 1, q 2 Aj(r i ) s.t. y(q 1 ) [y(r i 1 ), y(r i )) n y(q 2 ) (y(r i ), y(r i+1 )], 2 i k 1. 5 Furthr Rsrh Dirtions Givn point st P n th 2D monoton spnning grph o P tht hs th lst numr o gs prou in polynomil tim? Dos thr xist t pproximtion lgorithm, t < 2, or th k root y monoton minimum spnning grph prolm? Aknowlgmnt: I woul lik to thnk Prossor Antonios Symvonis or his vlul ontriution in vloping th rsults prsnt in this ppr.

CCCG 2018, Winnipg, Cn, August 8 10, 2018 Rrns [1] S. Almri, T. M. Chn, E. Grnt, A. Luiw, n V. Pthk. Sl-pprohing grphs. In W. Diimo n M. Ptrignni, itors, Grph Drwing - GD 2012, volum 7704 o LNCS, pgs 260 271. Springr, 2013. [2] N. Alon, Z. Füri, n M. Kthlski. Sprting pirs o points y stnr oxs. Eur. J. Com., 6(3):205 210, 1985. [3] P. Anglini. Monoton rwings o grphs with w irtions. In. Pross. Ltt., 120:16 22, 2017. [4] P. Anglini, E. Colsnt, G. Di. Bttist, F. Frti, n M. Ptrignni. Monoton rwings o grphs. J. Grph Algorithms Appl., 16(1):5 35, 2012. [5] P. Anglini, W. Diimo, S. Koourov, T. Mhliz, V. Roslli, A. Symvonis, n S. Wismth. Monoton rwings o grphs with ix ming. Algorithmi, 71(2):233 257, 2015. [6] Y. Bhoo, S. Durohr, S. Mhrpour, n D. Monl. Exploring inrsing-hor pths n trs. In J. Gumunsson n M. Smi, itors, Proings o th 29th Cnin Conrn on Computtionl Gomtry, CCCG 2017, pgs 19 24. Crlton Univrsity, Ottw, Ontrio, Cn, 2017. [7] J. L. Bntly n H. A. Murr. Eiint worsts t struturs or rng srhing. At In., 13:155 168, 1980. [8] N. Bonihon, P. Bos, P. Crmi, I. Kostitsyn, A. Luiw, n S. Vronshot. Gril tringultions n ngl-monoton grphs: Lol routing n rognition. In Y. Hu n M. Nöllnurg, itors, Grph Drwing n Ntwork Visuliztion - 24th Intrntionl Symposium, GD 2016, Athns, Gr, Sptmr 19-21, 2016, Rvis Slt Pprs, volum 9801 o LNCS, pgs 519 531. Springr, 2016. [9] M. Brg, O. Chong, M. J. vn Krvl, n M. H. Ovrmrs. Computtionl gomtry: lgorithms n pplitions. Springr, 3r ition, 2008. [10] H. R. Dhkori, F. Frti, n J. Gumunsson. Inrsing-hor grphs on point sts. J. Grph Algorithms Appl., 19(2):761 778, 2015. [11] M. R. Gry, D. S. Johnson, F. P. Prprt, n R. E. Trjn. Tringulting simpl polygon. In. Pross. Ltt., 7(4):175 179, 1978. [12] N. Guttmnn-Bk n R. Hssin. On two rstrit nstors tr prolms. In. Pross. Ltt., 110(14-15):570 575, 2010. [13] D. H n X. H. Optiml monoton rwings o trs. SIAM Journl on Disrt Mthmtis, 31(3):1867 1877, 2017. [14] M. Ihino n J. Sklnsky. Th rltiv nighorhoo grph or mix tur vrils. Pttrn Rognition, 18(2):161 167, 1985. [15] D. G. Lrmn n P. MMulln. Ars with inrsing hors. Mthmtil Proings o th Cmrig Philosophil Soity, 72:205 207, Sptmr 1972. [16] D. T. L n F. P. Prprt. Lotion o point in plnr suivision n its pplitions. SIAM J. Comput., 6(3):594 606, 1977. [17] G. Liott, A. Luiw, H. Mijr, n S. Whitsis. Th rtngl o inlun rwility prolm. Comput. Gom., 10(1):1 22, 1998. [18] A. Luiw n D. Monl. Angl-monoton grphs: Constrution n lol routing. CoRR, s/1801.06290, 2018. [19] A. Luiw n J. O Rourk. Angl-monoton pths in non-otus tringultions. In J. Gumunsson n M. Smi, itors, Proings o th 29th Cnin Conrn on Computtionl Gomtry, CCCG 2017, pgs 25 30. Crlton Univrsity, Ottw, Ontrio, Cn, 2017. [20] K. Mstks. Drwing Root Tr s Root y Monoton Minimum Spnning Tr. CoRR, s/1806.04720, 2018. [21] K. Mstks n A. Symvonis. On th onstrution o inrsing-hor grphs on onvx point sts. In 6th Int. Con. on Inormtion, Intllign, Systms n Applitions, IISA 2015, Coru, Gr, July 6-8, 2015, pgs 1 6. IEEE, 2015. [22] K. Mstks n A. Symvonis. Root uniorm monoton minimum spnning trs. In D. Fotkis, A. Pgourtzis, n V. Th. Pshos, itors, Algorithms n Complxity - 10th Intrntionl Conrn, CIAC 2017, Athns, Gr, My 24-26, 2017, Proings, volum 10236 o LNCS, pgs 405 417, 2017. Full Vrsion:rXiv:1607.03338v2, 2017. [23] M. Nöllnurg, R. Prutkin, n I. Ruttr. On sl-pprohing n inrsing-hor rwings o 3-onnt plnr grphs. Journl o Computtionl Gomtry, 7(1):47 69, 2016.

30 th Cnin Conrn on Computtionl Gomtry, 2018 [24] A. Oikonomou n A. Symvonis. Simpl ompt monoton tr rwings. In F. Frti n K.-L. M, itors, Grph Drwing n Ntwork Visuliztion - 25th Intrntionl Symposium, GD 2017, Boston, MA, USA, Sptmr 25-27, 2017, Rvis Slt Pprs, volum 10692 o LNCS, pgs 326 333. Springr, 2017. [25] M. H. Ovrmrs n D. Woo. On rtngulr visiility. J. Algorithms, 9(3):372 390, 1988. [26] G. Rot. Curvs with inrsing hors. Mthmtil Proings o th Cmrig Philosophil Soity, 115:1 12, Jnury 1994. [27] M. I. Shmos n D. Hoy. Closst-point prolms. In 16th Annul Symposium on Fountions o Computr Sin, Brkly, Cliorni, USA, Otor 13-15, 1975, pgs 151 162. IEEE Computr Soity, 1975.