the game of Zombies and Survivors on graphs Anthony Bonato Ryerson University SIAM DM 2016 Cops and Robbers and PursuitEvasion in Discrete Structures About the minisymposium Cops and Robbers ID: 550804
Download Presentation The PPT/PDF document "A probabilistic version of" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
A probabilistic version of the game of Zombies and Survivors on graphs
Anthony BonatoRyerson University
SIAM
DM
2016
Cops and Robbers and Pursuit-Evasion in Discrete StructuresSlide2About the
minisymposiumCops and Robbers is a trending topic in graph theorystructural, probabilistic, topological, algorithmic…pursuit-evasion/graph searchingthree sessions and 14 speakers: Mon, Wed, ThursSlide3Cops and Robbers
CC
RSlide4Cops and Robbers
CC
RSlide5Cops and Robbers
two-player game, perfect information, unit speed, played on vertices, move to neighbours or passminimum number of cops needed for a winning strategy to capture R in G = cop number of G, or c(G)Slide6
Meyniel’s Conjecture (85): c(G) = O(n1/2) if G is connectedcomputing c(G) ≤ k(Berrarducci, Intrigila
, 93), (Hahn,MacGillivray, 06), (B,Chiniforooshan, 09): polynomial time if
k
is fixed
(Kinnersley,15)
:
EXPTIME
-complete, if
k
not fixed
not know if it is in
NP
(Aigner,Fromme,84)
:
c(G)
≤
3
if
G
is planar
Schroeder’s conjecture (01)
:
c(G)
≤
g + 3
, if
g
is the genus of
GSlide7Variants
power of cops:traps, photo radar, walls, capture from distance, teleportation, tandem-win, lazy copspower of robber:speed: fast or infinite, invisible, decoys, barricades, damage, capture copsvertex pursuit/search/good guys vs bad guys games:Firefighting, watchman’s problem, eternal domination, seepage, robot vacuum, robot crawler, graph cleaning, Prisoner’s dilemma, Walker-Breaker, Angel and Devil, …Slide8Slide9
Zombies and SurvivorsSlide10
Zombies and SurvivorsSlide11
Zombie horde
Zombies and Survivors
u
p to
- 2
zombies on an induced path will never capture the survivor
Slide12
Zombies and Survivorsset of zombies, one survivorplayers move at alternate ticks of the clock, from vertex to vertex along edgeszombies choose their initial locations u.a.r.at each step the zombies move along a shortest path connected to the survivorif more than one such path, then they choose one u.a.r.zombies
win if one or more can eat the survivorland on the survivor’s vertexotherwise, survivor winsNB: zombies have no strategy!
Zombies and SurvivorsSlide13
(B,Mitsche,Perez-Gimenez,Pralat,16+)sk(G): probability survivor wins if k zombies play, assuming optimal playsk+1 (G) ≤ sk (G) for all
k, and sk(G) → 0 as k → ∞
zombie number
of
G
is
z(G) = min{k ≥ c(G):
s
k
(G) ≤ ½}
well-defined
z(G)
represents the minimum number of zombies such that the probability that they eat the survivor is
> ½
note that
c(G) ≤ z(G)
Z(G)
= z(G) / c(G):
cost of being undead
Zombies and SurvivorsSlide14Deterministic Zombies
deterministic version of the game: (Fitzpatrick, Howell, Messinger,Pike,16+) at the SIAM DM 2014 conference in Minneapolisin the deterministic version, the zombies choose their location, and can choose their geodesics if more than oneexample: where c(G) = 2 < z(G) = 3
bridged graphs, products, hypercubes (z(G) - c(G) → ∞)
Zombies and SurvivorsSlide15
Zombies and Survivors
c(G) = 2
with probability
all
k
zombies begin
outside
the cycle
implies:
a
nd so
G
n-5
leavesSlide16Zombie number of cycles
Theorem (BMPGP,16+)If n ≥ 27, then z(Cn) = 4, so Z(Cn) = 2.
Zombies and SurvivorsSlide17Sketch of proof
survivor wins iff the zombies are on a subpath of order at most - 2overcount these configurations:
distinguishing two zombies one at each end of the subpathk(k-1) ways to select the two zombies on subpath with r ≤
-
2
vertices
n
choices for position of the path
two zombies start in right place with probability
(1/n)
2
remaining zombies on
subpath
, with probability
(r/n)
k-2
Zombies and SurvivorsSlide18Sketch of proof
hence,k = 4 gives sk(Cn) ≤
½lower bound analogous
Zombies and SurvivorsSlide19
Zombies and Survivorsfinite projective plane P:two lines meet in a unique pointtwo points determine a unique lineexist 4 points, no line contains more than two of themq2+q+1
points; each line (point) contains (is incident with) q+1 points (lines)incidence graph (IG) of P:
bipartite graph
G(P)
with
red
nodes the points of
P
and
blue
nodes the lines of
P
a point is joined to a line if it is on that lineSlide20
ExampleZombies and SurvivorsFano plane
Heawood graphSlide21
Cop number of planesIG of projective planes: girth 6, (q+1)-regular, so have cop number ≥ q+1order 2(q2+q+1)
(Prałat,10): cop number = q+1Zombies and SurvivorsSlide22Zombies on planes
Theorem (BMPGP,16+)For large q,
).
Hence,
Z(
) = 2.
Zombies and SurvivorsSlide23
(Maamoun,Meyniel,87)
Theorem (
BMPGP,16+)
For large
n
,
).
Hence,
Z(
) = 4/3.
Zombies and SurvivorsSlide24Sketch of proof of lower bound
show a.a.s. survivor can escape
zombies
vertices
: binary bit strings
Chernoff’s
bound
:
zombies begin with an even number of ones in their binary string
this holds throughout the game, as zombies always move
the survivor, independently, always has an even distance to less than
n/3
zombies, and an odd distance to less than
n/3
zombies
Zombies and SurvivorsSlide25Sketch of proof of lower bound
survivor strategy: pick an arbitrary vertex distance at least 2 from all the zombiesnote that there are at most k(n+1) < 2n vertices at distance at most 1 from some zombieif after the zombie’s move, no zombie is distance 1 to the survivor, then the survivor stands stillotherwise, survivor must move to a safe vertexeach zombie distance
1 (< n/3) forbids a coordinate, and each zombie at distance 2 (< n/3) forbids two coordinates
so less than
n
coordinates forbidden, so survivor can move
Zombies and SurvivorsSlide26Cartesian grids
(Tosic, 87) c(G H) ≤ c(G) + c(H)Theorem (BMPGP,16+)
For n ≥ 2, z(Pn Pn
) = 2
,
so
Z(P
n
P
n
) =1
.
Zombies and SurvivorsSlide27Toroidal grids
Tn = Cn Cn(Neufeld, 90): c(
Tn) = 3Theorem (BMPGP,16+)Let
ω
=
ω
(n)
be a function tending to infinity with
n
.
Then
a.a.s
.
/ (
ω
log n).
Zombies and SurvivorsSlide28Toroidal grids, continued
despite the lower bound, no known upper bound is known for the zombie number of toroidal graphs!Zombies and SurvivorsSlide29
Necromancers, Zombies, and Survivorsset of j necromancers (i.e. cops, who can access all strategies) and k zombies play vs survivor(j,k)-necro-win: bad guys win with probability > 1/2
Theorem (BMPGPR,16+)For large m and n
, toroidal grids
C
m
C
n
are
(2,1)
-
necro
-win.
If
gcd
(
m,n
) ≤ 3
, then
C
m
C
n
is
(1,2)
-
necro
-win
.
Zombies and SurvivorsSlide30Problems
zombie number of toroidal gridupper bound?other grids? graph products?necromancer + k zombies: how large must k be to ensure win on square toroidal grids?random structures? binomial random graph, G(n,r), random regular…
Zombies and SurvivorsSlide31Deterministic case
(FHMP,16+) characterize zombie-win graphsconjecture:
lower bound tight; best upper bound:
n-1
how large can
z(G)/c(C)
be?
Zombies and SurvivorsSlide32
CanaDAM 2017Ryerson University, TorontoSlide33Math PhD program at Ryerson
starts Fall 2016emphasis on modelling, applicationsdiscrete math/graph theory is a fieldtaking applications in October for Fall 2017http://math.ryerson.ca/graduate/overview-phd.htmlZombies and Survivors