Sunteți pe pagina 1din 37

Ant Colony Systems

and the
Ant Algorithm

Department of Biomedical, Industrial and Human Factors Engineering


1
REAL ANT BEHAVIOR

Department of Biomedical, Industrial and Human Factors Engineering


2
REAL ANT BEHAVIOR

 Natural behavior of ants have inspired scientists to mimic insect


operational methods to solve real-life complex problems
 By observing ant behavior, scientists have begun to understand
their means of communication
 Ant-based behavioral patterns to address combinatorial
problems - first proposed by Marco Dorigo

 Ants secrete pheromone while


traveling from the nest to food,
and vice versa in order to
communicate with one another to
find the shortest path

Department of Biomedical, Industrial and Human Factors Engineering


3
EXPERIMENTAL STUDY OF ANTS

The more ants follow a trail, the more attractive that trail
becomes for being followed

NEST FOOD

Department of Biomedical, Industrial and Human Factors Engineering


4
ANT Behavior

The more ants follow a trail, the more attractive that trail
becomes for being followed

Department of Biomedical, Industrial and Human Factors Engineering


5
ANT Behavior

Even when the tracks are equal the behavior will encourage one
over the other--convergence (Deneubourg et al)

Department of Biomedical, Industrial and Human Factors Engineering


6
ROUTE SELECTION

 Ants are forced to decide whether they should go left or right, and
the choice that is made is a random decision
 Pheromone accumulation is faster on the shorter path
 The difference in pheromone content between the two paths over
time makes the ants choose the shorter path
 Positive feedback mechanism to arrive at the shortest route while
foraging
 Stygmergy or stigmergetic model of communication
 Different optimization problems have been explored using a
simulation of this real ant behavior

Department of Biomedical, Industrial and Human Factors Engineering


7
TRAVELING SALESMAN PROBLEM

Department of Biomedical, Industrial and Human Factors Engineering


8
PROBLEM DEFINITION

OBJECTIVE

Given a set of n cities, the Traveling


Salesman Problem requires a salesman
to find the shortest route between the
given cities and return to the starting
city, while keeping in mind that each city
can be visited only once

Department of Biomedical, Industrial and Human Factors Engineering


9
WHY IS TSP DIFFICULT TO SOLVE?

 Finding the best solution may entail an exhaustive search for all
combinations of cities. This can be prohibitive as “n” gets very
large
 Heuristics like a “greedy” route doesn’t guarantee optimal
solutions

a d a d
e e
b f g b f g
c c
h h

Department of Biomedical, Industrial and Human Factors Engineering


10
TSP Applications

• Lots of practical applications


• Routing such as in trucking, delivery, UAVs
• Manufacturing routing such as movement of parts along
manufacturing floor or the amount of solder on circuit board
• Network design such as determining the amount of cabling required
• Two main types
– Symmetric
– Asymmetric

Department of Biomedical, Industrial and Human Factors Engineering


11
General Formulation - Symmetric

∑ ∑ d x
ij
ij
åå
x j
,
i+ x
i j>i
i,
j ³ 2 "
i
å
i åå
ÎSj
Ïx
>S
j
i
,
j
<i j
+
j
i i
å
ÏSj
Î
>i
x
>S
ji
i,j³ 2 "
S'
S³3
xi,jÎB+
Department of Biomedical, Industrial and Human Factors Engineering
12
∑ ∑
General Formulation - Asymmetric

 d x
ij
ij
å xj
,
i =
i j≠i
1 "i
å
j
xi
,
j =1 "i
å

ÎÏx
SjS
i
i

j1
xi,jÎB+
"S 'S
³2
Department of Biomedical, Industrial and Human Factors Engineering
13
TSP Heuristics

• Variety of heuristics used to solve the TSP


• The TSP is not only theoretically difficult it is also difficult in practical
application since the tour breaking contraints get quite numerous
• As a result there have been a variety of methods proposed for the TSP
• Nearest Neighbor is a typical greedy approach

Department of Biomedical, Industrial and Human Factors Engineering


14
Simple Examples

4 1 3
2 2
6 9 6
2 2
5 2 7 2 8 2 2
1 2 2 1
6 1 2 3 4 5 6 7 8 9 10
6 1 Chicago 96 105 50 41 86 46 29 56 70
2 Dallas 78 49 94 21 64 63 41 37
3 Denver 60 84 61 54 86 76 51
4 KC 45 35 20 26 17 18
5 Minn 80 36 55 59 64
6 OK City 46 50 28 8
7 Omaha 45 37 30
8 St Louis 21 45
9 Sprngfld 25
10 Wichita

Department of Biomedical, Industrial and Human Factors Engineering


15
Nearest Neighbor Solution
Iteration Node Arc Cost Total
1 1
2 8 1,8 29
3 9 8,9 21 50
4 4 9,4 17 67
5 10 4,10 18 85
6 6 10,6 8 93
7 2 6,2 21 114
8 7 2,7 64 178
9 5 7,5 36 214
10 3 5,3 84 298
11 3,1 105 403

Department of Biomedical, Industrial and Human Factors Engineering


16
Larger TSP Example
500

450

400

350
Obj Fx:
min d
300

Way Point
250
Depot

200

150

100

50

0
0 50 100 150 200 250 300 350 400 450 500
Department of Biomedical, Industrial and Human Factors Engineering
17
Initial Order Solution
500

450

400
d=3138
350

300

Way Point
250
Depot

200

150

100

50

0
0 Department
50 100 of Biomedical,
150 200 Industrial
250 and
300 Human
350 Factors
400 Engineering
450 500
18
Nearest Neighbor Solution
500

450

400
d=2108
350

300

Way Point
250
Depot

200

150

100

50

0
0 Department
50 100 of Biomedical,
150 200 Industrial
250 and
300 Human
350 Factors
400 Engineering
450 500
19
Tabu Search Solution
500

450

400
d=1830
350

300

Way Point
250
Depot

200

150

100

50

0
0 Department
50 100 of Biomedical,
150 200 Industrial
250 and
300 Human
350 Factors
400 Engineering
450 500
20
THE ANT COLONY OPTIMIZATION
METAHEURISTIC

Department of Biomedical, Industrial and Human Factors Engineering


21
GOAL OF ACO HEURISTIC

 Artificial ants form a multi-agent system performing the


functions as observed in the real ant system
 Exploit stigmergistic communication

The ACO meta-heuristic relies on the co-operation of a


group of artificial ants to obtain a good solution to a
discrete optimization problem such as the TSP

 Artificial ants are mutants of a real ant system


 The resulting shortest route mapping determined by the agents
can be applied to the optimization problem

Department of Biomedical, Industrial and Human Factors Engineering


22
ACO CHARACTERISTICS

 Exploit a positive feedback mechanism


 Demonstrate a distributed computational architecture
 Exploit a global data structure that changes dynamically as
each ant transverses the route
 Has an element of distributed computation to it involving the
population of ants
 Involves probabilistic transitions among states or rather
between nodes

Department of Biomedical, Industrial and Human Factors Engineering


23
REAL vs. ARTIFICIAL ANTS

• Discrete time steps


• Memory Allocation
• Quality of Solution
• Time of Pheromone
deposition
• Distance Estimation

REAL ANT ARTIFICIAL ANT

Department of Biomedical, Industrial and Human Factors Engineering


24
FLOWCHART OF ACO

START ACO

Locate ants randomly Have the


in cities across the NO maximum YES
grid and store the Iterations been
current city performed
in a tabu list

Determine the shortest


Determine probabilistically NO STOP
tour till now and
as to which city to visit next update pheromone ACO

Move to next city and Have all YES


Record the length of
place this city in the cities been
tour and clear tabu list
tabu list visited

Department of Biomedical, Industrial and Human Factors Engineering


25
KEY PARAMETERS

 Trail intensity is given by value of τij which indicates the intensity of


the pheromone on the trail segment, (ij)
 Trail visibility is ηij = 1/dij
 The importance of the intensity in the probabilistic transition is α
 The importance of the visibility of the trail segment is β
 The trail persistence or evaporation rate is given as ρ
 Q is a constant and the amount of peromone laid on a trail
segment employed by an Ant; this amount may be modified in
various manners

Department of Biomedical, Industrial and Human Factors Engineering


26
PROBABILISTIC CITY SELECTION

 Helps determine the city to visit next while the ant is in a tour
 Determined by variables such as the pheromone content in an
edge (i,j) at time instant t, heuristic function of the desirability of
adding edge, and their control parameters

[ ][ ]
ì t ij (t ) a h ij b
ïï
p ij (t ) = í å [t il (t )] [h il ] if j Î J k (i )
k a b

ï lÎ J k (i )
ïî 0 f j Ï J k (i )

Department of Biomedical, Industrial and Human Factors Engineering


27
PHEROMONE UPDATING

 Using the tour length for the k-th Ant, Lk, the quantity of
pheromone added to each edge belonging to the completed tour is
given by
ì Q where edge (i, j ) Î Tk (t )
ï
Dt ijk (t ) = í Lk
ïî 0 if edge (i, j ) Ï Tk (t )

 The pheromone decay in each edge of a tour is given by

τ ij (t + 1) = (1 − ρ )τ ij (t ) + ∆τ ij (t )

Department of Biomedical, Industrial and Human Factors Engineering


28
ACTUALLY 3 ALGORITHMS

 The ant-cycle is the approach discussed so far


 Information is updated at the end of each tour as such function
of tour length
 The ant-density is an approach wherein the pheromone quantity Q
is deposited once the segment is transversed
 Pretty much a greedy approach (local information) and not
really providing relative information
 The ant-quantity is an approach wherein the pheromone quantity
Q/dij is deposited once the segment is transversed
 Also a greedy approach but providing some relative information
by scaling Q by the length of the segment

Department of Biomedical, Industrial and Human Factors Engineering


29
Consider the Case Studies in Papers

Department of Biomedical, Industrial and Human Factors Engineering


30
EXTENSIONS

 Found that communication among the ants via the intensity factor
is important; makes sense since it provides some global insight
 Found that a good number of ants is about equal to the number of
cities
 Found that the initial distribution of the ants among the cities does
not really matter
 Found that an elitist strategy in which the segments on the best
solution(s) is(are) continually reinforced work well so long as there
were not too few or too many elitist solutions

Department of Biomedical, Industrial and Human Factors Engineering


31
APPLICATIONS

Department of Biomedical, Industrial and Human Factors Engineering


32
APPLICATIONS

Department of Biomedical, Industrial and Human Factors Engineering


33
Population-Based Incremental Learning

 Lots of similarities to the ACO


 Actually inspired by genetic algorithms
 Generate members of a population randomly based on
probability of selection functions that are increased or
decreased based on the quality of past solutions involving the
member variables
 Once a population is generated, evaluate and then increase or
lower the probability used by the generating vector to
encourage better solutions.

Department of Biomedical, Industrial and Human Factors Engineering


34
Population-Based Incremental Learning

 Benefits
 Will converge to solutions under correct circumstances
 Efficient in terms of storage
 Computationally pretty cheap
 Involves learning
 Disadvantages
 Keeps primarily a local focus
 Cannot handle interdependence among parameters very
well
 Will need to involve penalty functions for constraints

Department of Biomedical, Industrial and Human Factors Engineering


35
Population-Based Incremental Learning

 A solution to overcome disadvantages proposed by Miagkiky


and Punch
 Combine reinforcement with population generation

Department of Biomedical, Industrial and Human Factors Engineering


36
QUESTIONS??

Department of Biomedical, Industrial and Human Factors Engineering


37

S-ar putea să vă placă și