Download - Atomic Routing Games on Maximum Congestion
![Page 1: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/1.jpg)
1
Atomic Routing Games on Maximum Congestion
Costas BuschDepartment of Computer Science
Louisiana State University
Collaborators: Rajgopal Kannan, LSUMalik Magdon-Ismail, RPI
![Page 2: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/2.jpg)
2
Introduction
Price of Stability
Price of Anarchy
Outline of Talk
Bicriteria Game
![Page 3: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/3.jpg)
3
Network RoutingEach player corresponds to a pair of source-destination
Objective is to select paths with small cost
![Page 4: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/4.jpg)
4
Main objective of each player is to minimize congestion: minimize maximum utilized edge
3 congestion C
iplayer
![Page 5: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/5.jpg)
5
A player may selfishly choose an alternativepath that minimizes congestion
CC 31 congestion
Congestion Games:
![Page 6: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/6.jpg)
6
Player cost function for routing :i
ii Cppc )(
p
Congestionof selected path
Social cost function for routing :
CpSC )(
p
Largest player cost
![Page 7: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/7.jpg)
We are interested in Nash Equilibriumswhere every player is locally optimal
Metrics of equilibrium quality:
p
Price of Stability
)()(min *pSCpSC
p
Price of Anarchy
)()(max *pSCpSC
p
*p is optimal coordinated routing with smallest social cost
![Page 8: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/8.jpg)
8
Results:• Price of Stability is 1
• Price of Anarchy is )log( nLO
Maximum allowed path length
![Page 9: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/9.jpg)
9
Introduction
Price of Stability
Price of Anarchy
Outline of Talk
Bicriteria Game
![Page 10: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/10.jpg)
10
We show:
• QoR games have Nash Equilibriums
(we define a potential function)
• The price of stability is 1
![Page 11: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/11.jpg)
11
],,,,,[)( 21 Nk mmmmpM
number of players with cost km k
Routing Vector
![Page 12: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/12.jpg)
12
Routing Vectors are ordered lexicographically
],,,[)( 21 NmmmpM
],,,[)( 21 NmmmpM
= = = =
],,,,,[)( 11 Nkk mmmmpM
],,,,,[)( 11 Nkk mmmmpM
< < = =
)()( pMpM
)()( pMpM )( pp
)( pp
![Page 13: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/13.jpg)
13
If player performs a greedy movetransforming routing to then:p p pp
iLemma:
Proof Idea:Show that the greedy move gives a lower order routing vector
![Page 14: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/14.jpg)
14
kk
ii Cppck )(ii Cppck )(
Player CostiBefore greedy move:After greedy move:
Since player cost decreases:
![Page 15: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/15.jpg)
15
],,,,,,,[)( 11 Nkkk mmmmmpM
Before greedy move player was counted herei
],,,,,,,[)( 11 Nkkk mmmmmpM
After greedy moveplayer is counted herei
![Page 16: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/16.jpg)
16
],,,,,,,[)( 11 rkkk mmmmmpM
],,,,,,,[)( 11 rkkk mmmmmpM
> ==No change
Definite Decrease
possibledecrease
possibleincreaseor decrease
Possible increase
>
END OF PROOF IDEA
![Page 17: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/17.jpg)
17
Existence of Nash Equilibriums
Greedy moves give lower order routings
Eventually a local minimum for every playeris reached which is a Nash Equilibrium
![Page 18: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/18.jpg)
18
minp
Price of Stability
Lowest order routing :
*min )( SCpSC
• Is a Nash Equilibrium
• Achieves optimal social cost
1)(Stability of Price *min
SCpSC
![Page 19: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/19.jpg)
19
Introduction
Price of Stability
Price of Anarchy
Outline of Talk
Bicriteria Game
![Page 20: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/20.jpg)
20
We show for any restricted QoR game:
Price of Anarchy = )log( nLO
![Page 21: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/21.jpg)
21
Path of player
Consider an arbitrary Nash Equilibriump
i
iCedgemaximum congestionin path
![Page 22: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/22.jpg)
22
must have an edge with congestion
Optimal path of player
In optimal routing :*p
i
iC1 iCC
)(111*)( ppcCCCppc iiii
**)( CpSC
Since otherwise:
![Page 23: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/23.jpg)
23
C
00
0
edges use that Paths: Congestion of Edges :ECE
In Nash Equilibrium social cost is: CSC
0 0
![Page 24: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/24.jpg)
24
C 1C1C
0 0
Edges in optimal paths of 0
![Page 25: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/25.jpg)
25
C 1C1C
0 01 1
11
1
edges use that Players:1 least at Congestion of Edges :
ECE
![Page 26: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/26.jpg)
26
C 1C1C 2C 2C2C2C
0 01 1
Edges in optimal paths of 1
![Page 27: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/27.jpg)
27
C *SC *SC 2C 2C2C
0 01 1
2C
2 2
22
2
edges use that Players:2 least at Congestion of Edges :
ECE
![Page 28: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/28.jpg)
28
In a similar way we can define:
jj
j
EjCE
edges use that Players: least at Congestion of Edges :
![Page 29: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/29.jpg)
29
,,,,
,,,,
3210
3210
EEEEWe obtain sequences:
There exist subsequence:110
110
,,,,,,,
s
ss EEEE
||2|| 1 jj EEWhere: ||2|| 1 ss EEand1sj
ns log
![Page 30: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/30.jpg)
30
||))1((|| 11 ss EsCL
|||| 1*
s
s
EC
Maximum edge utilization
Minimum edge utilization
LMaximum path length
)log(* nLOCC
ns log ||2|| 1 ss EEKnown relations
![Page 31: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/31.jpg)
31
Worst Case Scenario:
![Page 32: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/32.jpg)
32
Introduction
Price of Stability
Price of Anarchy
Outline of Talk
Bicriteria Game
![Page 33: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/33.jpg)
33
We consider Quality of Routing (QoR) congestion games where the pathsare partitioned into routing classes:
QQQ ,,, 21
)()()( 21 QSQSQS
With service costs:
Only paths in same routing class can causecongestion to each other
![Page 34: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/34.jpg)
34
An example:
•We can have routing classes)(lognO
•Each routing class contains paths with length in range
jQ]2,2( 1jj
12)( jjQS•Service cost:
•Each routing class uses a different wireless frequency channel
![Page 35: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/35.jpg)
35
Player cost function for routing :i
iii SCppc )(
p
Congestionof selected path
Cost of respectiverouting class
![Page 36: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/36.jpg)
36
Social cost function for routing :
SCpSC )(
p
Largest player cost
![Page 37: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/37.jpg)
37
Results:• Price of Stability is 1
• Price of Anarchy is
)log),(min( ** nSCO
![Page 38: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/38.jpg)
38
We consider restricted QoR games
For any path :p )(|| pSp
Path length Service Cost of path
![Page 39: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/39.jpg)
39
We show for any restricted QoR game:
Price of Anarchy = )log),(min( ** nSCO
![Page 40: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/40.jpg)
40
Path of player
Consider an arbitrary Nash Equilibriump
i
iCedgemaximum congestionin path
![Page 41: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/41.jpg)
41
must have an edge with congestion
Optimal path of player
In optimal routing :*p
i
iC
*SCC i
)(111 *** ppcSCCSSCSCcp iiiiiiii
***)( SCpSC
Since otherwise:
![Page 42: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/42.jpg)
42
C
00
0
edges use that Paths: Congestion of Edges :ECE
In Nash Equilibrium: SCSC
0 0
![Page 43: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/43.jpg)
43
C *SC *SC
0 0
Edges in optimal paths of 0
![Page 44: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/44.jpg)
44
C *SC *SC
0 01 1
11
*1
edges use that Players:least at Congestion of Edges :E
SCE
![Page 45: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/45.jpg)
45
C *SC *SC *2SC *2SC *2SC *2SC
0 01 1
Edges in optimal paths of 1
![Page 46: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/46.jpg)
46
C *SC *SC *2SC *2SC *2SC
0 01 1
*2SC
2 2
22
*2
edges use that Players:2least at Congestion of Edges :
ESCE
![Page 47: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/47.jpg)
47
In a similar way we can define:
jj
j
E
jSCE
edges use that Players:
least at Congestion of Edges : *
![Page 48: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/48.jpg)
48
,,,,
,,,,
3210
3210
EEEEWe obtain sequences:
There exist subsequence:110
110
,,,,,,,
s
ss EEEE
||2|| 1 jj EEWhere: ||2|| 1 ss EEand1sj
ns log
![Page 49: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/49.jpg)
49
||))1((|| 1*
1 ss ESsCL
|||| 1*
s
s
EC
Maximum edge utilization
Minimum edge utilization
*SLMaximum path length
)log( ** nSOCC
ns log ||2|| 1 ss EEKnown relations
![Page 50: Atomic Routing Games on Maximum Congestion](https://reader033.vdocuments.us/reader033/viewer/2022051518/56816100550346895dd041b5/html5/thumbnails/50.jpg)
50
)log( ** nSOCC
)log),(min( Anarchy of Price **** nSCOSCSC
We have:
By considering class service costs, we obtain: