![Page 1: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/1.jpg)
1/26
Spy Game on Graphs
Nathann Cohen1 Nıcolas A. Martins2 Fionn Mc Inerney3
Nicolas Nisse3 Stephane Perennes3 Rudini Sampaio2
1CNRS, Univ Paris Sud, LRI, Orsay, France
2Universidade Federal do Ceara, Fortaleza, Brazil
3Universite Cote d’Azur, Inria, CNRS, I3S, France
UAI, Santiago, Chile, November 30, 2017
Seminario del Doctorado en Ingenierıa de Sistemas Complejos
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 2: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/2.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 3: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/3.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 4: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/4.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1
1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 5: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/5.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 6: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/6.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1
1
1
1
1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 7: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/7.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 8: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/8.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1
1
1
1
1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 9: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/9.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 10: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/10.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1
1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 11: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/11.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1 11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 12: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/12.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 1
11 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 13: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/13.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 11
1 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 14: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/14.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 1
11 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 15: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/15.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 11
1 1
1
1 1
1
1 1
1
1 1 11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 16: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/16.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1
11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 17: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/17.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1 1
1 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 18: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/18.jpg)
2/26
Pursuit-Evasion Games
Mobile agents in a graph.
Turn-by-turn with 2 players.
Coordination for common goal, e.g.,
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1
1
1 1
1
1 1
1
1 1
1
1
11
Cops and Robbers (capture) (Quilliot,
1978 ; Nowakowski, Winkler, 1983)
1 1 1 1 1 1 1
1 1 1
1
1
1
1
1 1
1
1 111 1
1
1 1
1
1 1
1
1 1
11 1
1
1 1
Eternal Domination (protection)
(Goddard et al, 2005)
natural applications : coordination of mobile autonomous agents(Robotics, Network Security, Information Seeking...)
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 19: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/19.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 20: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/20.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 21: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/21.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 22: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/22.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
1
1
11 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 23: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/23.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 24: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/24.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
1
1
1
1 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 25: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/25.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 26: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/26.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
1
1 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 27: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/27.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 28: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/28.jpg)
3/26
Spy Game
Spy (1st) vs guards (2nd) in agraph G .
Start : Spy placed at a vertex.Then, guards placed.
Turn-by-turn : Spy traverses upto s ≥ 2 edges. Guards traverseup to 1 edge.
Goal : Spy wants to be at leastdistance d + 1 from all guards.
Ex : s = 2 and d = 1.
1 1 1 1
1
1 1 1
1
1
11
11 1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 29: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/29.jpg)
4/26
Guard Number : gns,d(G )
Definition
For all s ≥ 2, d ≥ 0 and a graph G , gns,d(G ) is the minimumnumber of guards guaranteed to win vs the spy.
1 1 1 1
1
1 11
gn2,1(G ) = 2
gns,1(G ) ≤ γ(G )
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 30: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/30.jpg)
4/26
Guard Number : gns,d(G )
Definition
For all s ≥ 2, d ≥ 0 and a graph G , gns,d(G ) is the minimumnumber of guards guaranteed to win vs the spy.
1 1 1 1
1
1 11
gn2,1(G ) = 2
gns,1(G ) ≤ γ(G )
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 31: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/31.jpg)
5/26
Our Results : Computing gn
Complexity
Calculating gns,d is NP-hard in general.
Tight bounds for paths
gns,d(Pn) =⌈
n2d+2+q
⌉where q = b 2d
s−1c.
Almost tight bounds for cycles⌈n+2q
2(d+q)+3
⌉≤ gns,d(Cn) ≤
⌈n+2q
2(d+q)+1
⌉where q = b 2d
s−1c.
Polynomial time Linear Program for trees
Can calculate gns,d(T ) and a corresp. strategy in polynomial time.
Grids
∃β > 0, s.t. Ω(n1+β) ≤ gns,d(Gn×n).
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 32: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/32.jpg)
6/26
Related Work
Cops vs robber (capture at a distance) (Bonato et al, 2010).
Cops vs fast robber (Fomin et al, 2010).
How many cops needed in an n × n grid ?
Eternal Domination (Goddard et al, 2005).
γm(m × n grid) ≤ dmn5 e+ O(m + n) (Lamprou et al, 2016).
γm(G ) = gns,d(G ) when s =∞ and d = 0.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 33: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/33.jpg)
6/26
Related Work
Cops vs robber (capture at a distance) (Bonato et al, 2010).
Cops vs fast robber (Fomin et al, 2010).
How many cops needed in an n × n grid ?
Eternal Domination (Goddard et al, 2005).
γm(m × n grid) ≤ dmn5 e+ O(m + n) (Lamprou et al, 2016).
γm(G ) = gns,d(G ) when s =∞ and d = 0.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 34: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/34.jpg)
6/26
Related Work
Cops vs robber (capture at a distance) (Bonato et al, 2010).
Cops vs fast robber (Fomin et al, 2010).
How many cops needed in an n × n grid ?
Eternal Domination (Goddard et al, 2005).
γm(m × n grid) ≤ dmn5 e+ O(m + n) (Lamprou et al, 2016).
γm(G ) = gns,d(G ) when s =∞ and d = 0.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 35: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/35.jpg)
6/26
Related Work
Cops vs robber (capture at a distance) (Bonato et al, 2010).
Cops vs fast robber (Fomin et al, 2010).
How many cops needed in an n × n grid ?
Eternal Domination (Goddard et al, 2005).
γm(m × n grid) ≤ dmn5 e+ O(m + n) (Lamprou et al, 2016).
γm(G ) = gns,d(G ) when s =∞ and d = 0.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 36: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/36.jpg)
6/26
Related Work
Cops vs robber (capture at a distance) (Bonato et al, 2010).
Cops vs fast robber (Fomin et al, 2010).
How many cops needed in an n × n grid ?
Eternal Domination (Goddard et al, 2005).
γm(m × n grid) ≤ dmn5 e+ O(m + n) (Lamprou et al, 2016).
γm(G ) = gns,d(G ) when s =∞ and d = 0.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 37: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/37.jpg)
6/26
Related Work
Cops vs robber (capture at a distance) (Bonato et al, 2010).
Cops vs fast robber (Fomin et al, 2010).
How many cops needed in an n × n grid ?
Eternal Domination (Goddard et al, 2005).
γm(m × n grid) ≤ dmn5 e+ O(m + n) (Lamprou et al, 2016).
γm(G ) = gns,d(G ) when s =∞ and d = 0.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 38: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/38.jpg)
7/26
Paths
: Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉
Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11 11 1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 39: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/39.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 11
1 11 11 1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 40: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/40.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 11 1
11 11 1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 41: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/41.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1
1 1
1 11 1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 42: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/42.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1
11
11 1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 43: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/43.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 1
1 1
1 1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 44: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/44.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11
11
1 111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 45: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/45.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11
11 1
111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 46: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/46.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11 1
1 1 1
11
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 47: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/47.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11 1
1
1
11
1
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 48: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/48.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11 11 1
111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 49: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/49.jpg)
7/26
Paths : Lower bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 1 1 1 1 1 1 1 1 1
1 1 11 11 1
111
gn3,1(P10) = 2
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 50: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/50.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11 111 11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 51: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/51.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 11
11 111 11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 52: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/52.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 11 1
1 111 11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 53: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/53.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1
11
111 11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 54: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/54.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 1
1 1
11 11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 55: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/55.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11
11
1 11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 56: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/56.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11 1
11
11 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 57: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/57.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11 11
1 1
1 11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 58: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/58.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11 111
11
11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 59: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/59.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11 111 1
1 1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 60: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/60.jpg)
8/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Ex : s = 3 and d = 1.
1 guard can protect subpath of 2d + 2 + b 2ds−1c vertices.
1 1 1 1 1
1 11 111 11
11
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 61: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/61.jpg)
9/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Each guard protects a subpath of 2d + 2 + b 2d
s−1c vertices.
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 62: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/62.jpg)
9/26
Paths : Upper bound
Theorem
For all s ≥ 2, d ≥ 0, and a path Pn on n vertices,
gns,d(Pn) =
⌈n
2d+2+b 2ds−1c
⌉Each guard protects a subpath of 2d + 2 + b 2d
s−1c vertices.
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 63: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/63.jpg)
10/26
Guard Strategies
Spy-Positional Strategy :
f : V k × V ⇒ V k (Unrestricted strategy)
ω : V ⇒ V k (Restricted or Spy-Positional strategy)
Guards’ positions depend only on position of spy.
Unique configuration for guards for each position of spy.
Zonal Strategy :
Divide graph into subgraphs & assign certain number of guards toeach.
Optimal strategy in paths uses both strategies above.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 64: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/64.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 65: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/65.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 66: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/66.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 67: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/67.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 68: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/68.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 69: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/69.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 70: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/70.jpg)
11/26
Cycles : Upper Bound Case 2d < s − 1
Theorem
For all s ≥ 2, d ≥ 0 s.t. 2d < s − 1,and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Ex : s = 6 and d = 0.
gn6,0(C12) = 4
1
11
1
1
1
1
11
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 71: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/71.jpg)
12/26
Cycles
Theorem
For all s ≥ 2, d ≥ 0 s.t. q = 0, and a cycle Cn on n vertices,
gns,d(Cn) =⌈
n2d+3
⌉.
Theorem
For all s ≥ 2, d ≥ 0 s.t. q 6= 0, and a cycle Cn on n vertices,⌈n+2q
2(d+q)+3
⌉≤ gns,d(Cn) ≤
⌈n+2q
2(d+q)+1
⌉.
Reminder : q =⌊
2ds−1
⌋.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 72: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/72.jpg)
13/26
Trees are Harder
Zonal strategy in Paths : 1 guard per subpath of 2d + 2 + b 2ds−1c
vertices.
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 73: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/73.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1 1
11
11
11 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 74: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/74.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1
1 1
11
11
11 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 75: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/75.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1 1
11
11
11 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 76: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/76.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1 1
11
11
11 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 77: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/77.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1
1 1
11
1
1
11 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 78: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/78.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 79: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/79.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
1
1
1
1 1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 80: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/80.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
1
1
11
1 111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 81: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/81.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
11
11 1
111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 82: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/82.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
11
1
1 1 1
11
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 83: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/83.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
11
1
1
1
11
1
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 84: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/84.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
11
11 1
111
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 85: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/85.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
1
1
11
11 1 1
11
1
1
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 86: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/86.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 1
11
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 87: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/87.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 11
1
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 88: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/88.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 11
1
1
1
1
1
1
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 89: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/89.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 11
1
1
1
1
1
1
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 90: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/90.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 11
1
11
1
1
1
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 91: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/91.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 11
1
11
1
1
1
1
1
1
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 92: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/92.jpg)
14/26
Trees are Harder
Can’t always divide tree into subtrees protected by a certainnumber of guards. No Zonal strategy in trees.
11 1
1111
111
111
1 1 1
1 1 1
1 1 1
1
1
1
1
1
1
1
1
1
1 1
1
11
11
11 1 11
1
11
1
11
1
11
Example of a tree T where s = 2, d = 1 and gn2,1(T ) = 4.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 93: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/93.jpg)
15/26
Fractional Version of the Game
Guards may be fractional entities ;movements rep. by flows.
Unchanged for spy. Total fractionof guards distance ≤ d from spymust be ≥ 1.
Linear program to compute optimalspy-positional fractional strategy.
Optimal fractional strategy ⇒optimal integral strategy in trees.
s = 2, d = 1.
1 1
11
1 1
d0.25 0.25
0.25 0.250.25
0.25
10.51
0.51
0.51
gn2,1(C6) = 2but
1.5 guards suffice.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 94: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/94.jpg)
15/26
Fractional Version of the Game
Guards may be fractional entities ;movements rep. by flows.
Unchanged for spy. Total fractionof guards distance ≤ d from spymust be ≥ 1.
Linear program to compute optimalspy-positional fractional strategy.
Optimal fractional strategy ⇒optimal integral strategy in trees.
s = 2, d = 1.
1 1
11
1 1
d0.25 0.25
0.25 0.250.25
0.25
10.51
0.51
0.51
gn2,1(C6) = 2but
1.5 guards suffice.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 95: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/95.jpg)
15/26
Fractional Version of the Game
Guards may be fractional entities ;movements rep. by flows.
Unchanged for spy. Total fractionof guards distance ≤ d from spymust be ≥ 1.
Linear program to compute optimalspy-positional fractional strategy.
Optimal fractional strategy ⇒optimal integral strategy in trees.
s = 2, d = 1.
1 1
11
1 1
d
0.25 0.25
0.25 0.250.25
0.25
10.51
0.51
0.51
gn2,1(C6) = 2but
1.5 guards suffice.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 96: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/96.jpg)
15/26
Fractional Version of the Game
Guards may be fractional entities ;movements rep. by flows.
Unchanged for spy. Total fractionof guards distance ≤ d from spymust be ≥ 1.
Linear program to compute optimalspy-positional fractional strategy.
Optimal fractional strategy ⇒optimal integral strategy in trees.
s = 2, d = 1.
1 1
11
1 1
d
0.25 0.25
0.25 0.250.25
0.25
10.51
0.51
0.51
gn2,1(C6) = 2but
1.5 guards suffice.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 97: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/97.jpg)
15/26
Fractional Version of the Game
Guards may be fractional entities ;movements rep. by flows.
Unchanged for spy. Total fractionof guards distance ≤ d from spymust be ≥ 1.
Linear program to compute optimalspy-positional fractional strategy.
Optimal fractional strategy ⇒optimal integral strategy in trees.
s = 2, d = 1.
1 1
11
1 1
d0.25 0.25
0.25 0.250.25
0.25
10.51
0.51
0.51
gn2,1(C6) = 2but
1.5 guards suffice.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 98: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/98.jpg)
15/26
Fractional Version of the Game
Guards may be fractional entities ;movements rep. by flows.
Unchanged for spy. Total fractionof guards distance ≤ d from spymust be ≥ 1.
Linear program to compute optimalspy-positional fractional strategy.
Optimal fractional strategy ⇒optimal integral strategy in trees.
s = 2, d = 1.
1 1
11
1 1
d0.25 0.25
0.25 0.250.25
0.25
10.51
0.51
0.51
gn2,1(C6) = 2but
1.5 guards suffice.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 99: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/99.jpg)
16/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
Theorem : Can transform optimal fractional strategy into optimalintegral strategy in polynomial time.
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
1
Fractional Conf.
1 1 1 1 1
1 1
1
0 0 0 1 0
54
94
24
1
Transition Phase
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
1
Integral Conf.
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 100: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/100.jpg)
16/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
Theorem : Can transform optimal fractional strategy into optimalintegral strategy in polynomial time.
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
1
Fractional Conf.
1 1 1 1 1
1 1
1
0 0 0 1 0
54
94
24
1
Transition Phase
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
1
Integral Conf.
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 101: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/101.jpg)
16/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
Theorem : Can transform optimal fractional strategy into optimalintegral strategy in polynomial time.
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
1
Fractional Conf.
1 1 1 1 1
1 1
1
0 0 0 1 0
54
94
24
1
Transition Phase
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
1
Integral Conf.
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 102: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/102.jpg)
16/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
Theorem : Can transform optimal fractional strategy into optimalintegral strategy in polynomial time.
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
1
Fractional Conf.
1 1 1 1 1
1 1
1
0 0 0 1 0
54
94
24
1
Transition Phase
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
1
Integral Conf.
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 103: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/103.jpg)
16/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
Theorem : Can transform optimal fractional strategy into optimalintegral strategy in polynomial time.
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
1
Fractional Conf.
1 1 1 1 1
1 1
1
0 0 0 1 0
54
94
24
1
Transition Phase
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
1
Integral Conf.
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 104: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/104.jpg)
16/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
Theorem : Can transform optimal fractional strategy into optimalintegral strategy in polynomial time.
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
1
Fractional Conf.
1 1 1 1 1
1 1
1
0 0 0 1 0
54
94
24
1
Transition Phase
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
1
Integral Conf.
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 105: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/105.jpg)
17/26
Opt. Fractional Strategy ⇒ Opt. Integral Strategy in Trees
1 1 1 1 1
1 1
1
24
14
34
54
24
24
34
24
Fractional Conf.
rounding
1 1 1 1 1
1 1
1
0 0 0 1 0
1 2
1
Integral Conf.
1 1 1 1 1
1 1
1
44
14
54
64
24
14
14
0
rounding
1 1 1 1 1
1 1
1
1 0 1 1 0
0 1
1
Tree’s protection and guards’ movements preserved.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 106: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/106.jpg)
18/26
Restricted Strategies
f : V k × V ⇒ V k (Unrestricted strategy)
ω : V ⇒ V k (Restricted or Spy-Positional strategy)
Guards’ positions depend only on position of spy.
Unique configuration for guards for each position of spy.
Theorem
Optimal fractional strategy ⇒ optimal fractional Spy-Positionalstrategy in trees.
Can calculate optimal Spy-positional fractional strategies withLinear Program in polynomial time.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 107: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/107.jpg)
18/26
Restricted Strategies
f : V k × V ⇒ V k (Unrestricted strategy)
ω : V ⇒ V k (Restricted or Spy-Positional strategy)
Guards’ positions depend only on position of spy.
Unique configuration for guards for each position of spy.
Theorem
Optimal fractional strategy ⇒ optimal fractional Spy-Positionalstrategy in trees.
Can calculate optimal Spy-positional fractional strategies withLinear Program in polynomial time.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 108: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/108.jpg)
18/26
Restricted Strategies
f : V k × V ⇒ V k (Unrestricted strategy)
ω : V ⇒ V k (Restricted or Spy-Positional strategy)
Guards’ positions depend only on position of spy.
Unique configuration for guards for each position of spy.
Theorem
Optimal fractional strategy ⇒ optimal fractional Spy-Positionalstrategy in trees.
Can calculate optimal Spy-positional fractional strategies withLinear Program in polynomial time.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 109: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/109.jpg)
18/26
Restricted Strategies
f : V k × V ⇒ V k (Unrestricted strategy)
ω : V ⇒ V k (Restricted or Spy-Positional strategy)
Guards’ positions depend only on position of spy.
Unique configuration for guards for each position of spy.
Theorem
Optimal fractional strategy ⇒ optimal fractional Spy-Positionalstrategy in trees.
Can calculate optimal Spy-positional fractional strategies withLinear Program in polynomial time.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 110: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/110.jpg)
19/26
Linear Program to Compute Spy-Positional Strategy
Spy-positional strategy : ω : V ⇒ V k
ωx ,u : quantity of guards on u when spy is on x .
fx ,x ′,u,u′ : quantity of guards that go from u to u′ when spy goesfrom x to x ′.
(1) Minimize∑v∈V
ωx0,v
Minimize number of guards.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 111: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/111.jpg)
19/26
Linear Program to Compute Spy-Positional Strategy
Spy-positional strategy : ω : V ⇒ V k
ωx ,u : quantity of guards on u when spy is on x .
fx ,x ′,u,u′ : quantity of guards that go from u to u′ when spy goesfrom x to x ′.
(1) Minimize∑v∈V
ωx0,v
Minimize number of guards.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 112: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/112.jpg)
20/26
Linear Program to Compute Spy-Positional Strategy
Spy-positional strategy : ω : V ⇒ V k
ωx ,u : quantity of guards on u when spy is on x .
fx ,x ′,u,u′ : quantity of guards that go from u to u′ when spy goesfrom x to x ′.
(2)∑
v∈Nd [x]
ωx ,v ≥ 1 ∀x ∈ V
Guarantees always at least 1 guard within distance d of spy.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 113: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/113.jpg)
21/26
Linear Program to Compute Spy-Positional Strategy
Spy-positional strategy : ω : V ⇒ V k
ωx ,u : quantity of guards on u when spy is on x .
fx ,x ′,u,u′ : quantity of guards that go from u to u′ when spy goesfrom x to x ′.
(3)∑
u′∈N[u]
fx ,x ′,u,u′ = ωx ,u ∀u ∈ V , x ′ ∈ Ns [x ]
(4)∑
u′∈N[u]
fx ,x ′,u′,u = ωx ′,u ∀u ∈ V , x ′ ∈ Ns [x ]
Guarantees validity of moves of guards when spy moves.
O(n4) real variables and constraints.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 114: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/114.jpg)
21/26
Linear Program to Compute Spy-Positional Strategy
Spy-positional strategy : ω : V ⇒ V k
ωx ,u : quantity of guards on u when spy is on x .
fx ,x ′,u,u′ : quantity of guards that go from u to u′ when spy goesfrom x to x ′.
(3)∑
u′∈N[u]
fx ,x ′,u,u′ = ωx ,u ∀u ∈ V , x ′ ∈ Ns [x ]
(4)∑
u′∈N[u]
fx ,x ′,u′,u = ωx ′,u ∀u ∈ V , x ′ ∈ Ns [x ]
Guarantees validity of moves of guards when spy moves.
O(n4) real variables and constraints.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 115: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/115.jpg)
22/26
Main Result : gn in Trees
Theorem
∀s > 1, d ≥ 0 and all trees T , gns,d(T ) and a correspondingstrategy can be calculated in polynomial time.
Idea of proof : Linear Program can compute opt. frac.Spy-positional strategy in polynomial time.
Run LP. From previous theorem, strategy is opt. frac.
Can transform opt. frac. into opt. int. in polynomial time.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 116: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/116.jpg)
23/26
Grids
Theorem
∃β > 0, s.t. ∀s > 1, d ≥ 0, Ω(n1+β) ≤ gns,d(Gn×n).
Idea of proof : Lower bound holds for fractional version.
Torus and grid have same order of number of guards.
Theorem
∃α ≥ log(3/2) ≈ 0.58, s.t. ∀s > 1, d ≥ 0,fgns,d(Gn×n) ≤ O(n2−α).
Idea of proof : Density function ω∗(v) = c(dist(v ,v0)+1)log3/2 for a
constant c > 0 satisfies LP.
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 117: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/117.jpg)
23/26
Grids
Theorem
∃β > 0, s.t. ∀s > 1, d ≥ 0, Ω(n1+β) ≤ gns,d(Gn×n).
Idea of proof : Lower bound holds for fractional version.
Torus and grid have same order of number of guards.
Theorem
∃α ≥ log(3/2) ≈ 0.58, s.t. ∀s > 1, d ≥ 0,fgns,d(Gn×n) ≤ O(n2−α).
Idea of proof : Density function ω∗(v) = c(dist(v ,v0)+1)log3/2 for a
constant c > 0 satisfies LP.
1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 118: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/118.jpg)
24/26
Distribution of Guards in the Torus for an optimalsymmetrical spy-positional strategy when n = 100,m = 100, s = 2 and d = 1
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 119: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/119.jpg)
25/26
Further Work
Determine gns,d(Gn×n).
Approximate gns,d(G ) in polynomial time in certain classes ofgraphs ?
Fractional approach applied to other combinatorial games.
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs
![Page 120: Spy Game on Graphs - · 1/26 Spy Game on Graphs Nathann Cohen1 N colas A. Martins2 Fionn Mc Inerney3 Nicolas Nisse3 St ephane P erennes3 Rudini Sampaio2 1CNRS, Univ Paris Sud, LRI,](https://reader033.vdocuments.site/reader033/viewer/2022052003/6015c08c343cce1c6a23e075/html5/thumbnails/120.jpg)
26/26
Gracias !
Cohen, Martins, Mc Inerney, Nisse, Perennes, Sampaio Spy Game on Graphs