Title: Approximation Algorithms:
1Approximation Algorithms
problems, techniques, and their use in game
theory
- Éva Tardos
- Cornell University
2What is approximation?
- Find solution for an optimization problem
guaranteed to have value close to the best
possible. - How close?
- additive error (rare)
- E.g., 3-coloring planar graphs is NP-complete,
but 4-coloring always possible - multiplicative error
- ?-approximation finds solution for an
optimization problem within an ? factor to the
best possible.
3Why approximate?
- NP-hard to find the true optimum
- Just too slow to do it exactly
- Decisions made on-line
- Decisions made by selfish players
4Outline of talk
- Techniques
- Greedy
- Local search
- LP techniques
- rounding
- Primal-dual
- Problems
- Disjoint paths
- Multi-way cut and labeling
- network design, facility location
- Relation to Games
- local search ? price of anarchy
- primal dual ? cost sharing
5Max disjoint paths problem
- Given graph G, n nodes, m edges, and source-sink
pairs. - Connect as many as possible via edge-disjoint
path.
t
s
t
s
t
s
s
t
6Greedy Algorithm
Greedily connect s-t pairs via disjoint paths, if
there is a free path using at most m½ edges
If there is no short path at all, take a single
long one.
7Greedy Algorithm
Theorem m½ approximation. Kleinberg96 Proo
f One path used can block m½ better paths
Essentially best possible m½-? lower bound
unless PNP by Guruswami, Khanna, Rajaraman,
Shepherd, Yannakakis99
8Disjoint pathsopen problem
- Connect as many as pairs possible via paths where
2 paths may share any edge
- Same practical motivation
- Best greedy algorithm n½ - (and also m1/3 -)
approximation Awerbuch, Azar, Plotkin93. - No lower bound
9Outline of talk
- Techniques
- Greedy
- Local search
- LP techniques
- rounding
- Primal-dual
- Problems
- Disjoint paths
- Multi-way cut and labeling
- network design, facility location
- Relation to Games
- local search ? Price of anarchy
- primal dual ? Cost sharing
10Multi-way Cut Problem
- Given
- a graph G (V,E)
- k terminals s1, , sk
- cost we for each edge e
- Goal Find a partition that separates terminals,
and minimizes the cost - ?e separated we
Separated edges
s3
s1
s4
s2
11Greedy Algorithm
- For each terminal in turn
- Find min cut separating si from other terminals
s1
s3
s4
s2
s1
s3
The next cut
s4
s2
12Theorem Greedy is a2-approximation
- Proof Each cut costs at most the optimums cut
Dahlhaus, Johnson, Papadimitriou, Seymour, and
Yannakakis94 - Cuts found by algorithm
s3
s1
Optimum partition
s4
s2
Selected cuts, cheaper than optimums cut,
but each edge in optimum is counted twice.
13Multi-way cuts extension
- Given
- graph G (V,E), we?0 for e ?E
- Labels L1,,k
- Lv ? L for each node v
- Objective Find a labeling of nodes such that
each node v assigned to a label in Lv and it
minimizes cost ?e separated we
part 3
part 1
Separated edges
part 2
part 4
14Example
s1
cheap
medium
expensive
s2
s3
- Does greedy work?
- For each terminal in turn
- Find min cut separating si from other terminals
15Greedy doesnt work
- Greedy
- For each terminal in turn
- Find min cut separating si from other terminals
- The first two cuts
s1
Remaining part not valid!
s2
s3
16Local search
- Boykov Veksler Zabih CVPR98 2-approximation
- Start with any valid labeling.
- 2. Repeat (until we are tired)
- Choose a color c.
- b. Find the optimal move where a subset of the
vertices can be recolored, but only with the
color c. - (We will call this a c-move.)
17A possible -move
Thm Boykov, Vekler, Zabih The best -move
can be found via an (s,t) min-cut
18Idea of the flow networkfor finding a -move
s all other terminals retain current color
G
sc change color to c
19Theorem local optimum is a 2-approximation
Partition found by algorithm
Cuts used by optimum
The parts in optimum each give a possible local
move
20Theorem local optimum is a 2-approximation
Partition found by algorithm
Possible move using the optimum
Changing partition does not help ? current cut
cheaper Sum over all colors Each edge in optimum
counted twice
21Metric labeling ? classification open problem
- Given
- graph G (V,E) we?0 for e ?E
- k labels L
- subsets of allowed labels Lv
- a metric d(.,.) on the labels.
- Objective Find labeling f(v)?Lv for each node v
to minimize - ?e(v,w) we d(f(v),f(w))
Best approximation known O(ln k ln ln k)
Kleinberg-T99
22Outline of talk
- Techniques
- Greedy
- Local search
- LP techniques
- rounding
- Primal-dual
- Problems
- Disjoint paths
- Multi-way cut and labeling
- network design, facility location
- Relation to Games
- local search ? Price of anarchy
- primal dual ? Cost sharing
23Using Linear Programs for multi-way cuts
- Using a linear program
- fractional cut
- ? probabilistic assignment of nodes to parts
Idea Find optimal fractional labeling via
linear programming
24Fractional Labeling
- Variables
- 0 ? xva ? 1 pnode, alabel in Lv
- xva ? fraction of label a
- used on node v
- Constraints
? xva 1
for all nodes v ? V
a?Lv
- each node is assigned to a label
- cost as a linear function of x
- ? we ½ ? xua - xva
e(u,v)
a?L
25From Fractional x to multi-way cut
- The Algorithm (Calinescu, Karloff, Rabani, 98,
Kleinberg-T,99) - While there are unassigned nodes
- select a label a at random
26The Algorithm (Cont.)
- While there are unassigned nodes
- select a label a at random
select 0 ? ? ? 1 at random assign all unassigned
nodes v to selected label a if xva ? ?
27Why Is This Choice Good?
- select 0 ? ? ? 1 at random
- assign all unassigned nodes v to selected label a
if xva ? ? - Note
- Probability of assigning node v to label a is
? xva - Probability of separating nodes u and v in this
iteration is ?xua xva
28From Fractional x to Multi-way cut (Cont.)
- Theorem Given a fractional x, we find multi-way
cut with expected - separation cost ? 2 (LP cost of x)
- Corollary if x is LP optimum . ?
2-approximation - Calinescu, Karloff, Rabani, 98
- 1.5 approximation for multi-way cut (does not
work for labeling) - Karger, Klein, Stein, Thorup, Young99 improved
bound ? 1.3438..
29Outline of talk
- Techniques
- Greedy
- Local search
- LP techniques
- rounding
- Primal-dual
- Problems
- Disjoint paths
- Multi-way cut and labeling
- network design, facility location
- Relation to Games
- local search ? Price of anarchy
- primal dual ? Cost sharing
30Metric Facility Location
- F is a set of facilities (servers).
- D is a set of clients.
- cij is the distance between any i and j in D ? F.
- Facility i in F has cost fi.
31Problem Statement
We need to 1) Pick a set S of facilities to
open. 2) Assign every client to an open
facility (a facility in S). Goal Minimize
cost of S ?p dist(p,S).
32What is known?
- All techniques can be used
- Clever greedy Jain, Mahdian, Saberi 02
- Local search starting with Korupolu, Plaxton,
and Rajaraman 98, can handle capacities - LP and rounding starting with Shmoys, T, Aardal
97 - Here primal-dual starting with Jain-Vazirani99
33What is the primal-dual method?
- Uses economic intuition from cost sharing
- For each requirement, like
- ?a?Lv xva 1, someone has to pay to make it
true - Uses ideas from linear programming
- dual LP and weak duality
- But does not solve linear programs
34Dual Problem Collect Fees
- Client p has a fee ap (cost-share)
- Goal collect as much as possible max ?p ap
- Fairness Do no overcharge for any subset A of
clients and any possible facility i we must have - ?p ?A ap dist(p,i) ? fi
amount client p would contribute to building
facility i.
35Exact cost-sharing
- All clients connected to a facility
- Cost share ap covers connection costs for each
client p - Costs are fair
- Cost fi of selecting a facility i is covered by
clients using it - ?p ap f(S) ?p dist(p,S) , and
- both facilities are fees are optimal
36Approximate cost-sharing
- Idea 1 each client starts unconnected, and with
fee ap0 - Then it starts raising what it is willing to pay
to get connected - Raise all shares evenly a
- Example
client
possible facility with its cost
37Primal-Dual Algorithm (1)
Its a 1 share could be used towards building a
connection to either facility
a 1
- Each client raises his fee a evenly what it is
willing to pay
38Primal-Dual Algorithm (2)
a 2
Starts contributing towards facility cost
- Each client raises evenly what it is willing to
pay
39Primal-Dual Algorithm (3)
a 3
Three clients contributing
- Each client raises evenly what it is willing to
pay
40Primal-Dual Algorithm (4)
4
a 3
Open facility
clients connected to open facility
- Open facility, when cost is covered by
contributions
41Primal-Dual Algorithm Trouble
4
i
j
a 3
p
Open facility
- Trouble
- one client p connected to facility i, but
contributes to also to facility j
42Primal-Dual Algorithm (5)
ghost
4
i
j
a 3
p
Open facility
- Close facility j will not open this facility.
- Will this cause trouble?
- Client p is close to both i and j ? facilities i
and j are at most 2a from each other.
43Primal-Dual Algorithm (6)
ghost
a 3
4
a 6
a 3
a 3
Open facility
no not need to pay more than 3
- Not yet connected clients raise their fee evenly
- Until all clients get connected
44Feasibility fairness ??
- ? All clients connected to a facility
- ? Cost share ap covers connection costs of
client p - ? Cost fi of opening a facility i is covered by
clients connected to it - ?? Are costs fair ??
45Are costs fair??
- a set of clients A, and any possible facility i
we have ?p ?A ap dist(p,i)? fi - Why? we open facility i if there is enough
contribution, and do not raise fees any further - But closed facilities are ignored! and may
violate fairness
46Are costs fair??
j
i
4
aq4
Closed facility, ignored
open facility
p
cause of closing
Fair till it reaches a ghost facility. Let aq
? aq be the fee till a ghost facility is reached
47Feasibility fairness ??
- ? All clients connected to a facility
- ? Cost share ap covers connection costs for
client p - ? Cost ap also covers cost of selected a
facilities - ? Costs ap are fair
- How much smaller is a ? a ??
48How much smaller is a ? a?
- q client met ghost facility j
- j became a ghost due to client p
j
i
4
q
p
- p stopped raising its share first
- ap ? aq ? aq
- Recall dist(i,j) ? 2 ap, so
- aq ? aq 2 ap ? 3aq
49Primal-dual approximation
- The algorithm is a 3-approximation
algorithm for the facility location problem - Jain-Vazirani99, Mettu-Plaxton00
- Proof
- Fairness of the ap fees ?
- ?p ap ? min cost max ? min
- cost-recovery
- f(S) ?p dist(p,S) ?p ap
- a ? 3aq
- 3-approximation algorithm
50Outline of talk
- Techniques
- Greedy
- Local search
- LP techniques
- rounding
- Primal-dual
- Problems
- Disjoint paths
- Multi-way cut and labeling
- network design, facility location
- Relation to Games
- primal dual ? Cost sharing
- local search ? Price of anarchy
51primal dual ? Cost sharing
- Dual variables ap are natural cost-shares
- Recall
- fair no set is overcharged
- core allocation
- ?p ?Aap dist(p,i) ? fi for all A and i.
- Chardaire98 Goemans-Skutella00 strong
connection between core cost-allocation and
linear programming dual solutions - See also Shapley67, Bondareva63 for other games
52Primal-Dual ? Cost-sharing
- Primal dual for each requirement someone
willing to pay to make it true - Cost-sharing only players can have shares.
- Not all requirements are naturally associated
with individual players. - Real players need to share the cost.
53primal dual ? Cost sharing
- Fair ? no subset is overcharged
- Stronger desirable property population monotone
(cross-monotone) - Extra clients do not increase cost-shares.
- Spanning-tree game Kent and Skorin-Kapov96 and
Jain Vazirani01 - Facility location, single source rent-or-buy
Pal-T02
54Local search (for facility location)
- Local search simple search steps to improve
objective - add(s) adds new facility s
- delete(t) closes open facility t
- swap(s,t) replaces open facility s by a new
facility t - Key to approximation bound
- How bad can be a local optima?
- 3-approximation Charikar, Guha00
55Local search ? Price of anarchy in games
- Price of anarchy facilities are operated by
separate selfish agents - Agents open/close facilities when it benefits
their own objective. - Agents best response dynamic
- Simple local steps analogous to local search.
- Price of anarchy
- How bad can be a stable state?
- 2-approximation in a related maximization game
Vetta02
56Conclusions for approximation
- Greedy, Local search
- clever greedy/local steps can lead to great
results - Primal-dual algorithms
- Elegant combinatorial methods
- Based on linear programming ideas, but fast,
avoids explicitly solving large linear programs - Linear programming
- very powerful tool, but slow to solve
- Interesting connections to issues in game theory