• No results found

Ants coony optimiztion problem

N/A
N/A
Protected

Academic year: 2020

Share "Ants coony optimiztion problem"

Copied!
70
0
0

Loading.... (view fulltext now)

Full text

(1)

Ant Colony Optimization:

an introduction

(2)

2

Outline

1. Biological inspiration of ACO

2. Solving NP-hard combinatorial problems 3. The ACO metaheuristic

(3)

Outline

1. Biological inspiration of ACO

2. Solving NP-hard combinatorial problems 3. The ACO metaheuristic

(4)

4

(5)

Ant colonies

• Distributed systems of social insects • Consist of simple individuals

(6)

6

Ant Cooperation

Stigmergy – indirect communication

between individuals (ants)

(7)

Denebourg’s double bridge

experiments

• Studied Argentine ants I. humilis

(8)

8

(9)

Double bridge experiments:

equal lengths (2)

• Run for a number of trials

(10)

10

(11)

Double bridge experiments:

different lengths (2)

(12)

12

Outline

1. Biological inspiration of ACO

2. Solving NP-hard combinatorial problems

3. The ACO metaheuristic

(13)
(14)

14

Combinatorial optimization

• Find values of discrete variables

(15)

Combinatorial optimization

Π = (S, f, Ω) – problem instance • S – set of candidate solutionsf – objective function

• Ω – set of constraints

• – set of feasible solutions (with respect to Ω)

• Find globally optimal feasible solution s*

(16)

16

NP-hard combinatorial problems

• Cannot be exactly solved in polynomial time

Approximate methods – generate

near-optimal solutions in reasonable time • No formal theoretical guarantees

(17)

Approximate methods

(18)

18

Constructive algorithms

• Add components to solution incrementally • Example – greedy heuristics:

(19)

Local search

• Explore neighborhoods of complete solutions

• Improve current solution by local changes

(20)

20

What is a metaheuristic?

• A set of algorithmic concepts

(21)

Examples of metaheuristics

• Simulated annealing • Tabu search

• Iterated local search

• Evolutionary computation

Ant colony optimization

(22)

22

Outline

1. Biological inspiration of ACO

2. Solving NP-hard combinatorial problems

3. The ACO metaheuristic

(23)
(24)

24

ACO metaheuristic

• A colony of artificial ants cooperate in finding good solutions

• Each ant – simple agent

• Ants communicate indirectly using

(25)

Combinatorial optimization

problem mapping (1)

• Combinatorial problem (S, f, Ω(t)) • Ω(t) – time-dependent constraints

 Example – dynamic problems

• Goal – find globally optimal feasible solution s*

• Minimization problem

(26)

26

Combinatorial optimization

problem mapping (2)

C = {c1, c2, …, cNc} – finite set of

components

States of the problem:

X = {x = <ci, cj, …, ch, …>, |x| < n < +∞} • Set of candidate solutions:

(27)

Combinatorial optimization

problem mapping (3)

• Set of feasible states:

• We can complete into a solution satisfying Ω(t)

• Non-empty set of optimal solutions:

X

X

~

X x  ~

S

X

(28)

28

Combinatorial optimization

problem mapping (4)

S*

X

~

S

X

• X – states

• S – candidate solutions • – feasible states

S* – optimal solutions

(29)

Combinatorial optimization

problem mapping (5)

• Cost g(s, t) for each

• In most cases – g(s, t) ≡ f(s, t)

GC = (C, L) – completely connected graph • C – set of components

• L – edges fully connecting the components (connections)

Gconstruction graph

S

s

S

s

~

(30)

30

Combinatorial optimization

problem mapping (last

)

(31)

Construction graph

• Each component ci or connection lij have associated:

 heuristic information

(32)

32

Heuristic information

A priori information about the problem • Does not depend on the ants

• On components ciηi

• On connections lij – ηij

• Meaning:

(33)

Pheromone trail

• Long-term memory about the entire search process

• On components ciτi

• On connections lij – τij

(34)

34

Artificial ant (1)

• Stochastic constructive procedure • Builds solutions by moving on GC

• Has finite memory for:

– Implementing constraints Ω(t) – Evaluating solutions

(35)

Artificial ant (2)

• Has a start state x

• Has termination conditions ek

• From state xr moves to a node from the neighborhood – Nk(x

r)

(36)

36

Artificial ant (3)

• Selects a move with a probabilistic rule depending on:

 Pheromone trails and heuristic information of neighbor components and connections

 Memory

(37)

Artificial ant (4)

• Can update pheromone on visited components (nodes)

• and connections (edges) • Ants act:

 Concurrently

(38)

38

The ACO metaheuristic

While not doStop():

(39)

ConstructAntSolutions

• A colony of ants build a set of solutions • Solutions are evaluated using the

(40)

40

UpdatePheromones

• Two opposite mechanisms:

 Pheromone deposit

(41)

UpdatePheromones: pheromone

deposit

• Ants increase pheromone values on

visited components and/or connections • Increases probability to select visited

(42)

42

UpdatePheromones: pheromone

evaporation

• Decrease pheromone trails on all

components/connections by a same value • Forgetting – avoid rapid convergence to

(43)

DaemonActions

• Optional centralized actions, e.g.:

 Local optimization

(44)

44

ACO applications

• Traveling salesman

• Quadratic assignment • Graph coloring

• Multiple knapsack • Set covering

• Maximum clique • Bin packing

(45)

Outline

1. Biological inspiration of ACO

2. Solving NP-hard combinatorial problems 3. The ACO metaheuristic

(46)

46

(47)

Traveling salesman problem

N – set of nodes (cities), |N| = n

• A – set of arcs, fully connecting N

• Weighted graph G = (N, A)

• Each arc has a weight dijdistance • Problem:

(48)

48

TSP: construction graph

• Identical to the problem graph • C = N

L = A

(49)

TSP: constraints

• All cities have to be visited

• Each city – only once

(50)

50

TSP: pheromone trails

(51)

TSP: heuristic information

ηij = 1 / dij

(52)

52

TSP: solution construction

• Select random start city

(53)

ACO algorithms for TSP

• Ant System

• Elitist Ant System

• Rank-based Ant System • Ant Colony System

(54)

54

Ant System: Pheromone

initialization

• Pheromone initialization

Τij = m / Cnn,

where:

m – number of ants

Cnn – path length of nearest-neighbor

(55)

Ant System: Tour construction

• Ant k is located in city i

• is the neighborhood of city i

• Probability to go to city :

il il ij ij k ij

p

(56)

56

Tour construction:

comprehension

α = 0 – greedy algorithm

• β = 0 – only pheromone is at work

quickly leads to stagnation

k i N l il il ij ij k ij

p

(57)

Ant System: update pheromone

trails – evaporation

Evaporation for all connections∀(i, j) ∈ L:

τij ← (1 – ρ) τij,

ρ ∈[0, 1] – evaporation rate

(58)

58

Ant System: update pheromone

trails – deposit

Tk – path of ant k

Ck – length of path Tk

Ants deposit pheromone on visited arcs:

 

i

j

L

m

k

k ij ij

ij

,

,

1

 

     otherwise T j i

Ck k

(59)

Elitist Ant System

• Best-so-far ant deposits pheromone on each iteration: best ij m k k ij ij

ij

e

1

 

otherwise

T

j

i

(60)

60

Rank-based Ant System

• Rank all ants

(61)

MAX-MIN Ant System

1. Only iteration-best or best-so-far ant deposits pheromone

(62)

62

Ant Colony System

• Differs from Ant System in three points:

 More aggressive tour construction rule

 Only best ant evaporates and deposits pheromone

(63)

Ant Colony System

 



otherwise

System

Ant

like

q

q

if

j

l Nik il il

,

,

max

arg

0

1.Tour Construction

(64)

64

(65)

State of the art in TSP

• CONCORDE

http://www.tsp.gatech.edu/concorde.html • Solved an instance of 85900 cities

(66)

66

Current ACO research activity

(67)

Further reading

• M. Dorigo, T. Stützle. Ant Colony Optimization. MIT Press, 2004.

(68)

68

Next time…

(69)

Thank you!

Any questions?

(70)

References

Related documents

Coffee leaf temperatures were between 1-7° C lower in the shaded agroforestry system (coffee with Inga densiflora shade trees, with percentage shade cover between 25 and 50) than

After completing this lesson, the participant will be able to: • list the required steps and files to prepare a DMO procedure • prepare the archives for a DMO procedure.

In Tunisian oil oxygenated monoterpenes were found to be the major components of Artemisia herba-alba oil extracted from aerial parts.. In the essential oil of

Schematic description of the proposed search method: Instead of closest point searching from the root of the tree to the leaves that contain the data points, a pointer to the leaves

Accordingly, we had four major objectives: (i) to identify the optimum soil moisture for the occurrence of maximum GHG fluxes, (ii) to assess the impact of sand addition to the

Abbreviations: hepg2, human liver hepatocellular carcinoma cell line; ga, gambogic acid; cdTe QDs, cadmium–tellurium quantum dots; ga-cdTe, nanocomposites based on ga and cdTe

The IHOPE initiative has received support from several sources, including: the National Center for Ecological Analysis and Synthesis (NCEAS), the Stockholm Resilience Center, the