Martin Grtschel Institute of Mathematics, Technische Universitt Berlin (TUB)

DFG-Research Center Mathematics for key technologies (FZT 86)

Konrad-Zuse-Zentrum fr Informationstechnik Berlin (ZIB)
groetschel@zib.de http://www.zib.de/groetschel
Mathematical Challenges
in Telecommunication
a Tutorial
Martin Grtschel

IMA workshop on
Network Management and Design
Minneapolis, MN, April 6, 2003
Martin
Grtschel
2
Comment
This is an abbreviated version of my
presentation.
The three films I showed are not included.
I also deleted many of the pictures and some of
the tables:
in some cases I do not have permission to publish
them since they contain some confidential material or
information not for public use,
in other cases I am not sure of the copy right status
(and do not have the time and energy to check it).
Martin
Grtschel
3
The ZIB Telecom Group

Andreas Bley
Andreas Eisenbltter
Martin Grtschel
Thorsten Koch
Arie Koster
Roland Wessly
Adrian Zymolka

ZIB Associates

Manfred Brandt
Sven Krumke
Frank Lutz
Diana Poensgen
Jrg Rambau
and more:
Clyde Monma
(BellCore, ...)
Mechthild Opperud
(ZIB, Telenor)
Dimitris Alevras (ZIB, IBM)
Christoph Helmberg (Chemnitz)
ZIB Telecom Team
Martin
Grtschel
4
ZIB Partners from Industry
Bell Communications Research
Telenor (Norwegian Telecom)
E-Plus (acquired by KPN in 01/2002)
DFN-Verein
Bosch Telekom (bought by Marconi)
Siemens
Austria Telekom (controlled by Italia Telecom)
T-Systems Nova (T-Systems, Deutsche Telekom)
KPN
Telecel-Vodafone

Atesio (ZIB spin-off company)

Martin
Grtschel
5
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future
Martin
Grtschel
6
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future
Martin
Grtschel
7
What is the Telecom Problem?
Design excellent technical devices
and a robust network that survives
all kinds of failures and organize
the traffic such that high quality
telecommunication between
very many individual units at
many locations is feasible
at low cost!
Speech
Data
Video
Etc.
Martin
Grtschel
8
What is the Telecom Problem?
Design excellent technical devices
and a robust network that survives
all kinds of failures and organize
the traffic such that high quality
telecommunication between
very many individual units at
many locations is feasible
at low cost!
This problem is
too general
to be solved in
one step.
Approach in Practice:
Decompose whenever possible
Look at a hierarchy of problems
Address the individual problems one by one
Recompose to find a good global solution
Martin
Grtschel
9
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization
as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future
Martin
Grtschel
10
Cell Phones and Mathematics
Designing mobile phones
Task partitioning
Chip design (VLSI)
Component design
Computational logic
Combinatorial
optimization
Differential algebraic
equations
Producing Mobile Phones
Production facility layout
Control of CNC machines
Control of robots
Lot sizing
Scheduling
Logistics
Operations research
Linear and integer programming
Combinatorial optimization
Ordinary differential equations
Marketing and Distributing Mobiles
Financial mathematics
Transportation optimization
cell
phone
picture
Chip Design
CMOS layout for
four-transistor
static-memory
cell
CMOS layout for two
four-transistor
static-memory cells.
Compacted CMOS
layout for two
four-transistor
static-memory cells.
Placement
Routing
Compactification
Schematic for four-transistor
static-memory cell
ZIB-Film 9:01 9:41
Martin
Grtschel
12
Design and Production of ICs and PCBs
Integrated Circuit (IC)
Printed Circuit Board (PCB)
Problems: Logic Design, Physical Design
Correctness, Simulation, Placement of
Components, Routing, Drilling,...
Production and Mathematics:
Examples
CNC Machine for 2D and 3D
cutting and welding
(IXION ULM 804)
Sequencing of Tasks
and Optimization of Moves
Mounting Devices
Minimizing Production Time
via TSP or IP
Printed Circuit
Boards
Optimization of
Manufacturing
SMD
Martin
Grtschel
14
Drilling 2103 holes into a PCB
Significant Improvements
via TSP

(Padberg & Rinaldi)
Martin
Grtschel
15
Siemens Problem
before
after
printed circuit board da1
Martin
Grtschel
16
Siemens Problem
before after
printed circuit board da4
Martin
Grtschel
17
Mobile Phone Production Line
Fujitsu Nasu plant
Martin
Grtschel
18
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future
Martin
Grtschel
19
Network Components
Design, Production, Marketing, Distribution:
Similar math problems as for mobile phones
Fiber (and other) cables
Antennas and Transceivers
Base stations (BTSs)
Base Station Controllers (BSCs)
Mobile Switching Centers (MSCs)
and more...
Martin
Grtschel
20
Component Cables
Martin
Grtschel
21
Component Antennas
Martin
Grtschel
22
Component Base Station
Nokia MetroSite
Nokia UltraSite
Component
Mobile
Switching
Center:

Example of
an MSC Plan
Martin
Grtschel
24
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future

Martin
Grtschel
25
Network Design: Tasks to be solved
Some Examples
Locating the sites for antennas (TRXs) and
base transceiver stations (BTSs)
Assignment of frequencies to antennas
Cryptography and error correcting encoding for
wireless communication
Clustering BTSs
Locating base station controllers (BSCs)
Connecting BTSs to BSCs
Martin
Grtschel
26
Network Design: Tasks to be solved
Some Examples (continued)
Locating Mobile Switching Centers (MSCs)
Clustering BSCs and Connecting BSCs to MSCs
Designing the BSC network (BSS) and the
MSC network (NSS or core network)
Topology of the network
Capacity of the links and components
Routing of the demand
Survivability in failure situations
Most of these problems turn out to be
Combinatorial Optimization or
Mixed Integer Programming Problems
Martin
Grtschel
27
Connecting Mobiles: Whats up?
BSC
MSC
BSC
BSC
BSC
BSC
BSC
BSC
MSC
MSC
MSC
MSC
BTS
Martin
Grtschel
28
Connecting Computers or
other Devices
The mathematical
problems are similar
but not identical
Martin
Grtschel
29
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization
as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future

Martin
Grtschel
30
F
A
P

F
i
l
m
Martin
Grtschel
31
Antennas & Interference
x
x
antenna
backbone
network
x
x
x
x
site
x
x
cell
co- & adjacent
channel
interference
cell
Martin
Grtschel
32
Interference
Level of interference depends on
distance between transmitters
geographical position
power of the signals
direction in which signals are transmitted
weather conditions
assigned frequencies
co-channel interference
adjacent-channel interference
Martin
Grtschel
33
Separation
Parts of the spectrum forbidden
at some locations:
government regulations,
agreements with operators in
neighboring regions,
requirements military forces, etc.
Site
Blocked channels
Frequencies assigned to the same
location (site) have to be separated
Martin
Grtschel
34
Frequency Planning Problem
Find an assignment of frequencies to
transmitters that satisfies
all separation constraints
all blocked channels requirements
and either
avoids interference at all
or
minimizes the (total/maximum) interference level
Martin
Grtschel
35
Minimum Interference
Frequency Assignment Problem
Integer Linear Program:
{ }
min
. . 1
1 , ( )
1 ,
1 , 1
, , 0,1
e e
e
+
= e
+ s e <
+ s + e e
+ s + e =
e

co ad
v
co co ad ad
vw vw vw vw
vw E vw E
vf
f F
d
vf wg
co co
vf wf vw v w
ad ad
vf wg vw
co ad
vf vw vw
c z c z
s t x v V
x x vw E f g d vw
x x z vw E f F F
x x z vw E f g
x z z
Martin
Grtschel
36
A Glance at some Instances
I
n
s
t
a
n
c
e
|
V
|
d
e
n
s
i
t
y

[
%
]
m
i
n
i
m
u
m

d
e
g
r
e
e
a
v
e
r
a
g
e

d
e
g
r
e
e
m
a
x
i
m
u
m

d
e
g
r
e
e
d
i
a
m
e
t
e
r
c
l
i
q
u
e

n
u
m
b
e
r
k 267 56,8 2 151,0 238 3 69
B-0-E-20 1876 13,7 40 257,7 779 5 81
f 2786 4,5 3 135,0 453 12 69
h 4240 5,9 11 249,0 561 10 130
E-Plus Project
Martin
Grtschel
37
Region Berlin - Dresden
2877
carriers

50 channels

Interference
reduction:
83.6%
Martin
Grtschel
38
Region Karlsruhe
2877
Carriers

75 channels

Interference
Reduction:
83.9 %
Martin
Grtschel
39
The UMTS Radio Interface
Completely new story
see talk by Andreas Eisenbltter on Monday
Martin
Grtschel
40
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future

Martin
Grtschel
41
G-WiN Data
G-WiN = Gigabit-Wissenschafts-Netz of the DFN-Verein
Internet access of all German universities
and research institutions

Locations to be connected: 750
Data volume in summer 2000: 220 Terabytes/month
Expected data volume in 2004: 10.000 Terabytes/month

Clustering (to design a hierarchical network):
10 nodes in Level 1a 261 nodes eligible for
20 nodes in Level 1b Level 1
All other nodes in Level 2
Martin
Grtschel
42
G-WiN Problem

Select the 10 nodes of Level 1a.
Select the 20 nodes of Level 1b.

Each Level 1a node has to be linked to two Level 1b nodes.
Link every Level 2 node to one Level 1 node.

Design a Level 1a Network such that
Topology is survivable (2-node connected)
Edge capacities are sufficient (also in failure situations)
Shortest path routing (OSPF) leads to balanced
capacity use (objective in network update)
The whole network should be stable for the future.
The overall cost should be as low as possible.

Martin
Grtschel
43
Potential node locations for the
3-Level Network of the G-WIN
Red nodes are potential
level 1 nodes
Cost:
Connection between nodes
Capacity of the nodes
Blue nodes are all
remaining nodes
Martin
Grtschel
44
Demand distribution
The demand scales with the
height of each red line
Aim
Select backbone nodes and
connect all non-backbone nodes to
a backbone node
such that the
overall network cost is minimal
(access+backbone cost)
Martin
Grtschel
45
G-WiN Location Problem: Data
set of locations
set of potential Level 1a locations (subset of )
set of possible configurations at
location in Level 1a
For , and :
connection costs from to
tr
p
p
ip
i
V
Z V
K
p
i V p Z k K
w i p
d
=
=
=
e e e
=
= affic demand at location
capacity of location in configuration
costs at location in configuration
1 if location is connected to (else 0)
1 if configuration is used at loc
k
p
k
p
ip
k
p
i
c p k
w p k
x i p
z k
=
=
=
= ation (else 0) p
Martin
Grtschel
46
G-WiN Location/Clustering Problem
min
1
1
e e e e
+
=
s
=
=

p
k k
ip ip p p
p Z i V p Z k K
ip
p
k k
i ip p p
i k
k
p
k
k
p
p
w x w z
x
d x c z
z
z const
Each location i must be connected to a Level 1 node
Capacity at p must be large enough
Only one configuration at each Location 1 node
All variables are 0/1.
# of Level 1a nodes
Martin
Grtschel
47
Solution: Hierarchy & Backbone
Martin
Grtschel
48
G-WiN Location Problem:
Solution Statistics
The DFN problem leads to ~100.000 0/1-variables.
Typical computational experience:
Optimal solution via CPLEX in a few seconds!
A very related problem at Telekom Austria has
~300.000 0/1-variables plus some continuous variables
and capacity constraints.
Computational experience (before problem specific fine tuning):
10% gap after 6 h of CPLEX computation,
60% gap after simplification
(dropping certain capacities).
Martin
Grtschel
49
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future

Martin
Grtschel
50
Re-Optimization of
Signaling Transfer Points
Telecommunication companies maintain a signaling
network (in adition to their communication transport
network). This is used for management tasks such as:
Basic call setup or tear down
Wireless roaming
Mobile subscriber authentication
Call forwarding
Number display
SMS messages
Etc.
Martin
Grtschel
51
Signaling Transfer Point (STP)
CCD
CCD
CCD
CCD CCD
CCD
CCD
CCD
CCD
Link-Sets
STP
Cluster
Cluster
Cluster
CCLK
CCD=Common Channel Distributors, CCLK=Common Channel Link Controllers
CCD=routing unit, CCLK=interface card
Martin
Grtschel
52
STP Problem description
Target
Assign each link to a CCD/CCLK
Constraints
At most 50% of the links in a linkset can be assigned to a
single cluster
Number of CCLKs in a cluster is restricted
Objective
Balance load of CCDs
Martin
Grtschel
53
STP Mathematical model
Data
C set of CCDs j
L set of links i
Di demand of link i
P set of link-sets
Q set of clusters
Lp subset of links in link-set p
Cq subset of CCDs in cluster q
cq #CCLKs in cluster q
Variables

{ }
0,1 , , e e e
ij
x i L j C
1 =
ij
x
if and only if
link i is assigned to CCD j
Martin
Grtschel
54
STP Mathematical model
{ }
2
min
1
,
0,1
p
p q
q
ij
j C
i ij
i L
i ij
i L
ij
i L j C
ij q
i L j C
ij
L
y z
x i L
D x y j C
D x z j C
x p P q Q
x c q Q
x
e
e
e
e e
e e

= e
s e
> e
(
s e e
(
(
s e
e

Assign each link

Upper bound of CCD-load
Diversification
Lower bound of CCD-load
CCLK-bound
Integrality
Min load difference
Martin
Grtschel
55
STP current (former) solution
Minimum: 186 Maximum: 404 Load difference: 218
0
50
100
150
200
250
300
350
400
450
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
CCD
T
r
a
f
f
i
c

L
o
a
d
Martin
Grtschel
56
STP Optimal solution
Minimum: 280 Maximum: 283 Load difference: 3
0
50
100
150
200
250
300
350
400
450
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
CCD
T
r
a
f
f
i
c

L
o
a
d
Martin
Grtschel
57
STP Practical difficulty
Problem: 311 rearrangements are necessary
to migrate to the optimal solution
Reformulation with new objective
Find a best solution with a
restricted number of changes
Martin
Grtschel
58
STP Reformulated Model
{ }
*
, ( )
2
min
1
,
0,1
p
p q
q
ij
j C
i ij
i L
i ij
i L
ij
i L j C
ij q
i L j C
ij
ij
i L
j C j j i
L
y z
x i L
D x y j C
D x z j C
x p P q Q
x c q Q
x
x B
e
e
e
e e
e e
e
e =

= e
s e
> e
(
s e e
(
(
s e
e
s

Assign each link
Upper bound of CCD-load
Diversification
Lower bound of CCD-load
CCLK-bound
Integrality
Restricted number of changes!
Min load difference
Martin
Grtschel
59
STP Alternative Model
{ }
*
, ( )
2
min
1
,
0,1
p
p q
q
ij
i L j C j j i
ij
j C
i ij
i L
i ij
i L
ij
i L j C
ij q
i L j C
ij
L
x
x i L
D x y j C
D x z j C
x p P q Q
x c q Q
x
y z D
e e =
e
e
e
e e
e e
= e
s e
> e
(
s e e
(
(
s e
e
s

Assign each link

Upper bound of CCD-load
Diversification
Lower bound of CCD-load
CCLK-bound
Integrality
Restricted load difference
Min # changes
Martin
Grtschel
60
STP New Solutions
0
50
100
150
200
250
300
350
400
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
CCD
T
r
a
f
f
i
c

L
o
a
d
White: D=50, (alternative)
Minimum: 257 Maximum: 307
D=Load difference: 50 Number of changes: 12
Orange: B=8, (reformulated)
Minimum: 249 Maximum: 339
Load difference: 90 Number of changes=B: 8
Martin
Grtschel
61
STP Experimental results
Max changes 0 5 10 15 20
Load differences 218 129 71 33 14
1 hour application of CPLEX MIP-Solver for each case
Martin
Grtschel
62
STP - Conclusions
It is possible to achieve
85%
of the optimal improvement with less than
5%
of the changes necessary to obtain a load
balance optimal solution !
Martin
Grtschel
63
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future

Martin
Grtschel
64
Network Optimization Film
ZIB Film 1994, 9:40
Martin
Grtschel
65
Network Optimization
Networks
Capacities Requirements
Cost
Martin
Grtschel
66
What needs to be planned?
Topology
Capacities
Routing
Failure Handling (Survivability)

IP Routing
Node Equipment Planning
Optimizing Optical Links and Switches

DISCNET: A Network Planning Tool
(Dimensioning Survivable Capacitated NETworks)
Atesio ZIB Spin Off
Martin
Grtschel
67
The Network Design Problem
200
65
258
134
30
42
Dsseldorf
Frankfurt
Berlin
Hamburg
Mnchen
Communication Demands
Dsseldorf
Frankfurt
Berlin
Hamburg
Mnchen
Potential topology
&
Capacities
Martin
Grtschel
68
Capacities
PDH
2 Mbit/s
34 Mbit/s
140 Mbit/s
SDH
155 Mbit/s
622 Mbit/s
2,4 Gbit/s
... WDM (n x STM-N)
Two capacity models : Discrete Finite Capacities
Divisible Capacities
(P)SDH=(poly)synchronous digital hierarchy
WDM=Wavelength Division Multiplexer
STM-N=Synchronous Transport Modul with N STM-1 Frames
Martin
Grtschel
69
Survivability
H B
D
F
M
120
H B
D
F
M
30
30
60
Diversification
route node-disjoint
Path restoration
reroute affected demands
(or p% of all affected demands)
60
60
H B
D
F
M
H B
D
F
M
H B
D
F
M
60
60
Reservation
reroute all demands
(or p% of all demands)
120
H B
D
F
M
H B
D
F
M
H B
D
F
M
60
60
Martin
Grtschel
70
Model: Data & Variables
S s e Operating states:
s
uv s
P D uv S s P e e e , ,
0 ) ( > P f
s
uv
Path variables:
Capacity variables:
E ee } 1 , 0 { ) , ( e t e x
Supply Graph: G=(V,E)
+
eZ
0
e
C
e
T
e e
C C s s
1
s
uv
P P e
Valid Paths:
Demand Graph: H=(V,D)
D u v d
u v
e ,
D u v
u v
e ,
D u v l
u v
e ,
D u v
u v
e , o
Martin
Grtschel
71
Model: Capacities
Capacity variables : e e E, t = 1, ..., Te
{0,1}
t
e
x e
Cost function :
1
min
e
T
t t
e e
e E t
k x
e =

Capacity constraints : e e E
0 1
1 0
e
T
e e e
x x x = > > > >
0
e
T
t t
e e e
t
y c x
=
=

Martin
Grtschel
72
Model: Routings
Path variables :
Capacity constraints :
Demand constraints :
, ,
s
s uv
s S uv D P e e eP
( ) 0
s
uv
f P >
0
0
:
( )
uv
e uv
uv D
P e P
y f P
e
eP e
>

e E e
0
0
( )
uv
uv uv
P
d f P
eP
=

uv D e
Path length restriction
Martin
Grtschel
73
Model: Survivability (one example)
Path restoration reroute affected demands
H B
D
F
M
120
60
60
H B
D
F
M
for all seS, uveDs
for all seS, eeEs
60
60
H B
D
F
M
Martin
Grtschel
74
Mathematical Model
topology decisison
capacity decisions
normal operation routing
component failure routing
s
uv s
P D uv S s P e e e , , 0 ) ( > P f
s
uv
) (
:
0
0
P f y
D uv P e P
uv e
uv

e e P e
> E ee

P e
=
0
) (
0
uv
P
uv uv
P f d D uv e
1
min
e
T
t t
e e
e E t
k x
e =

e
T t E e , , 1 , = e
, 1, ,
e
e E t T e =
{0,1}
t
e
x e
t
e
t
e
x x >
1

=
=
e
T
t
t
e
t
e e
x c y
0
E ee
LP-based Methods
Feasible
integer
solutions
LP-based
relaxation
Convex
hull
Objective
function
Cutting
planes
Feasible
integer
solutions
LP-based
relaxation
Convex
hull
Objective
function
Cutting
planes
Flow chart
Optimal
solution
Initialize
LP-relaxation
Run
heuristics
Solve
LP-relaxation
Separation
algorithms
Augment
LP-relaxation
Solve feasibility
problem
Separation
algorithms
Inequalities?
No
Yes
x variables
integer?
Yes
No
Feasible
routings?
Yes
No
Polyhedral combinatorics
Valid inequalities (facets)
Separation algorithms
Heuristics
Feasibility of a capacity vector
LP-based approach:
Martin
Grtschel
77
Finding a Feasible Solution?
Heuristics
Local search
Simulated Annealing
Genetic algorithms
...
Nodes Edges Demands Routing-Paths
15 46 78 > 150 x 10e6
36 107 79 > 500 x 10e9
36 123 123 > 2 x 10e12
Manipulation of
Routings
Topology
Capacities
Problem Sizes
Martin
Grtschel
78
How much to save?
Real scenario
163 nodes
227 edges
561 demands
34% potential savings!
==
> hundred million dollars
PhD Thesis:
http://www.zib.de/wessaely
wessaely@atesio.de
Martin
Grtschel
79
Contents
1. Telecommunication: The General Problem
2. The Problem Hierarchy: Cell Phones and Mathematics
3. The Problem Hierarchy: Network Components and Math
4. Network Design: Tasks to be solved
Addressing Special Issues:
5. Frequency Assignment
6. Locating the Nodes of a Network
7. Balancing the Load of Signaling Transfer Points
8. Integrated Topology, Capacity, and Routing
Optimization as well as Survivability Planning
9. Planning IP Networks
10. Optical Networks
11. Summary and Future
Martin
Grtschel
80
Comment
9. Planning IP Networks
10. Optical Networks
11. Summary and Future


The lecture ended after about 100 minutes. The last
three topics above were not covered.

Martin
Grtschel
81
Summary
Telecommunication Problems such as

Frequency Assignment
Locating the Nodes of a Network Optimally
Balancing the Load of Signaling Transfer Points
Integrated Topology, Capacity, and Routing Optimization
as well as Survivability Planning
Planning IP Networks
Optical Network Design
and many others

can be succesfully attacked with optimization techniques.
Martin
Grtschel
82
Summary
The mathematical programming approach

Helps understanding the problems arising
Makes much faster and more reliable planning possible
Allows considering variations and scenario analysis
Allows the comparision of different technologies
Yields feasible solutions
Produces much cheaper solutions than traditional
planning techniques
Helps evaluating the quality of a network.

There is still a lot to be done, e.g.,
for the really important problems,
optimal solutions are way out of reach!
Martin
Grtschel
83
The Mathematical Challenges
Finding the right ballance between
flexibility and controlability of future networks
Controlling such a flexible network
Handling the huge complexity
Integrating new services easily
Guaranteeing quality

Finding appropriate Mathematical Models
Finding appropriate solution techniques (exact,
approximate , interactive, quality guaranteed)
Martin Grtschel Institute of Mathematics, Technische Universitt Berlin (TUB)
DFG-Research Center Mathematics for key technologies (FZT 86)
Konrad-Zuse-Zentrum fr Informationstechnik Berlin (ZIB)
groetschel@zib.de http://www.zib.de/groetschel
Mathematical Challenges
in Telecommunication
The End
Martin Grtschel
IMA workshop on
Network Management and Design
Minneapolis, MN, April 6, 2003