swarm intelligencce

8/2/2019 Swarm Intelligencce

1/30


2/30

Some examples are group foraging of social insects, cooperative transportation, division of labour,

nest-building of social insects, collective sorting and clustering

1.1 Analogies in IT and social insects

a) Distributed system of interacting autonomous agents

b) Goals: performance optimization and robustness

c) Self-organized control and cooperation (decentralized)

d) Division of labour and distributed task allocation

e) Indirect interactions

1.2 The 3 step process

1) Identification of analogies: in swarm biology and IT systems

2) Understanding: computer modelling of realistic swarm biology

3) Engineering: model simplification and tuning for IT applications

2. TAXONOMY OF SWARM INTELLINGENCE

2


3/30

Swarm intelligence has a marked multidisciplinary character since systems with the above mentioned

characteristics can be observed in a variety of domains. Research in swarm intelligence can be classified according todifferent criteria.

Natural vs. Artificial: It is customary to divide swarm intelligence research into two areas according to

the nature of the systems under analysis. We speak therefore of natural swarm intelligence research, where

biological systems are studied; and of artificial swarm intelligence, where human artifacts are studied.

Scientific vs. Engineering: An alternative and somehow more informative classification of swarm

intelligence research can be given based on the goals that are pursued: we can identify a scientific and

an engineering stream. The goal of the scientific stream is to model swarm intelligence systems and to single out

and understand the mechanisms that allow a system as a whole to behave in a coordinated way as a result of local

individual-individual and individual-environment interactions. On the other hand, the goal of the engineering

stream is to exploit the understanding developed by the scientific stream in order to design systems that are able to

solve problems of practical relevance.

The two dichotomies natural/artificial and scientific/engineering are orthogonal: although the typical

scientific investigation concerns natural systems and the typical engineering application concerns the development

of an artificial system, a number of swarm intelligence studies have been performed with swarms of robots for

validating mathematical models of biological systems. These studies are of a merely speculative nature and

definitely belong in the scientific stream of swarm intelligence. On the other hand, one could influence or modify the

behavior of the individuals in a biological swarm so that a new swarm-level behavior emerges that is somehow

functional to the solution of some task of practical interest. In this case, although the system at hand is a natural

one, the goals pursued are definitely those of an engineering application. In the following, an example is given for

each of the four possible cases.

2.1 Natural/Scientific: Foraging Behaviour of Ants

In a now classic experiment done in 1990, Deneubourg and his group showed that, when given the choice

3


4/30

between two paths of different length joining the nest to a food source, a colony of ants has a high probability to

collectively choose the shorter one. Deneubourg has shown that this behavior can be explained via a simple

probabilistic model in which each ant decides where to go by taking random decisions based on the intensity of

pheromone perceived on the ground, the pheromone being deposited by the ants while moving from the nest to the

food source and back.

2.2 Artificial/Scientific: Clustering by a Swarm of Robots

Several ant species cluster corpses to form cemeteries. Deneubourg et al. (1991) were among the first to

propose a distributed probabilistic model to explain this clustering behavior. In their model, ants pick up and drop

items with probabilities that depend on information on corpse density which is locally available to the ants. Beckers

et al. (1994) have programmed a group of robots to implement a similar clustering behavior demonstrating in this

way one of the first swarm intelligence scientific oriented studies in which artificial agents were used.

2.3 Natural/Engineering: Exploitation of collective behaviors of animal socities

A possible development of swarm intelligence is the controlled exploitation of the collective behavior of

animal societies. No example is available in this area of swarm intelligence although some promising research is

currently in progress: For example, in the Leurre project, small insect-like robots are used as lures to influence the

behavior of a group of cockroaches. The technology developed within this project could be applied to various

domains including agriculture and cattle breeding.

2.4 Artificial/Engineering:Swarm-based Data Analysis

Engineers have used the models of the clustering behavior of ants as an inspiration for designing data

mining algorithms . A seminal work in this direction was undertaken by Lumer and Faieta in 1994. They defined an

4
http://leurre.ulb.ac.be/index2.htmlhttp://leurre.ulb.ac.be/index2.htmlhttp://leurre.ulb.ac.be/index2.htmlhttp://www.scholarpedia.org/article/Data_mininghttp://www.scholarpedia.org/article/Data_mininghttp://www.scholarpedia.org/article/Data_mininghttp://www.scholarpedia.org/article/Data_mininghttp://www.scholarpedia.org/article/Algorithmhttp://leurre.ulb.ac.be/index2.htmlhttp://www.scholarpedia.org/article/Data_mininghttp://www.scholarpedia.org/article/Data_mininghttp://www.scholarpedia.org/article/Algorithm


5/30

artificial environment in which artificial ants pick up and drop data items with probabilities that are governed by the

similarities of other data items already present in their neighborhood. The same algorithm has also been used for

solving combinatorial optimization problems reformulated as clustering problems (Bonabeau et al. 1999).

3. PROPERTIES OF SWARM INTELLIGENCE SYSTEM

The typical swarm intelligence system has the following properties:

5


6/30

1) It is composed of many individuals;

2) The individuals are relatively homogeneous (i.e., they are either all identical or they belong to a few typologies);

3) The interactions among the individuals are based on simple behavioral rules that exploit only local information

that the individuals exchange directly or via the environment (stigmergy);

4) The overall behaviour of the system results from the interactions of individuals with each other and with their

environment, that is, the group behavior self-organizes.

The characterizing property of a swarm intelligence system is its ability to act in a coordinated way without

the presence of a coordinator or of an external controller. Many examples can be observed in nature of swarms that

perform some collective behavior without any individual controlling the group, or being aware of the overall group

behavior. Notwithstanding the lack of individuals in charge of the group, the swarm as a whole can show an

intelligent behavior. This is the result of the interaction of spatially neighboring individuals that act on the basis of

simple rules.

Most often, the behavior of each individual of the swarm is described in probabilistic terms: Each individual

has a stochastic behavior that depends on his local perception of the neighborhood.

Because of the above properties, it is possible to design swarm intelligence system that are scalable, parallel,

and fault tolerant.

Scalability means that a system can maintain its function while increasing its size without the need to

redefine the way its parts interact. Because in a swarm intelligence system interactions involve only neighboring

individuals, the number of interactions tends not to grow with the overall number of individuals in the swarm: each

individual's behavior is only loosely influenced by the swarm dimension. In artificial systems, scalability is

interesting because a scalable system can increase its performance by simply increasing its size, without the need for

any reprogramming.

Parallel action is possible in swarm intelligence systems because individuals composing the swarm can

perform different actions in different places at the same time. In artificial systems, parallel action is desirable

because it can help to make the system more flexible, that is, capable to self-organize in teams that take care

6


7/30

simultaneously of different aspects of a complex task.

Fault tolerance is an inherent property of swarm intelligence systems due to the decentralized, self-

organized nature of their control structures. Because the system is composed of many interchangeable individuals

and none of them is in charge of controlling the overall system behavior, a failing individual can be easily dismissed

and substituted by another one that is fully functioning.

4. STUDIES AND APPLICATIONS OF SWARMINTELLIGENCE

This section briefly presents a few examples of scientific and engineering swarm intelligencestudies.

7


8/30

Clustering Behavior of Ants

Ants build cemeteries by collecting dead bodies into a single place in the nest. They also organize the

spatial disposition of larvae into clusters with the younger, smaller larvae in the cluster center and the older ones

at its periphery. This clustering behavior has motivated a number of scientific studies. Scientists have built simple

probabilistic models of these behaviors and have tested them in simulation (Bonabeau et al. 1999). The basic

models state that an unloaded ant has a probability to pick up a corpse or a larva that is inversely proportional

to their locally perceived density, while the probability that a loaded ant has to drop the carried item is

proportional to the local density of similar items. This model has been validated against experimental data

obtained with real ants. In the taxonomy this is an example of natural/scientific swarm intelligence system.

Nest Building Behavior of Wasps and Termites

Wasps build nests with a highly complex internal structure that is well beyond the cognitive capabilities of a

single wasp. Termites build nests whose dimensions (they can reach many meters of diameter and height)

are enormous when compared to a single individual, which can measure as little as a few millimeters.

Scientists have been studying the coordination mechanisms that allow the construction of these structures and

have proposed probabilistic models exploiting stigmergic communication to explain the insects' behavior.

Some of these models have been implemented in computer programs and used to produce simulated structures

structures that recall the morphology of the real nests (Bonabeau et al. 1999). In the taxonomy this is an example

of natural/scientific swarm intelligence system.

Flocking and Schooling in Birds and Fish

Flocking and schooling are examples of highly coordinated group behaviors exhibited by large groups of birds

and fish. Scientists have shown that these elegant swarm-level behaviors can be understood as the result of a

self-organized process where no leader is in charge and each individual bases its movement decisions solely

on locally available information: the distance, perceived speed, and direction of movement of neighbours.

These studies have inspired a number of computer simulations (of which Reynolds' Boids simulation program

was the first one) that are now used in the computer graphics industry for the realistic reproduction of flocking

in movies and computer games. In the taxonomy these are examples respectively of natural/scientific and

artificial/engineering swarm intelligence systems.

8
http://www.red3d.com/cwr/boids/http://www.red3d.com/cwr/boids/http://www.red3d.com/cwr/boids/


9/30


10/30

agentslocate optimal solutions by moving through a parameter space representing all possible

solutions. Real ants lay down pheromones directing each other to resources while exploring their

environment. The simulated 'ants' similarly record their positions and the quality of their solutions,

so that in later simulation iterations more ants locate better solutions. [4] One variation on this approach is

the bees algorithm , which is more analogous to the foraging patterns of the honey bee .

Ant colony optimization (Dorigo, Maniezzo and Colorni 1991; Dorigo and Sttzle 2004) is a

population-based metaheuristic that can be used to find approximate solutions to difficult optimization

problems. It is inspired by the above-described foraging behavior of ant colonies. In ant colony

optimization (ACO), a set of software agents called "artificial ants" search for good solutions to a given

optimization problem transformed into the problem of finding the minimum cost path on a

weighted graph . The artificial ants incrementally build solutions by moving on the graph. The solution

construction process is stochastic and is biased by a pheromone model, that is, a set of parameters

associated with graph components (either nodes or edges) the values of which are modified at runtime

by the ants. ACO has been applied successfully to many classical combinatorial optimization problems,

as well as to discrete optimization problems that have stochastic and/or dynamic components. Examples

are the application to routing in communication networks (see also the Swarm

based Network Management section below) and to stochastic version of well-known combinatorial

optimization problem, such as the probabilistic traveling salesman problem. Moreover, ACO has been

extended so that it can be used to solve continuous and mixed-variable optimization problems (Socha

and Dorigo in press). Ant colony optimization is probably the most successful example of

artificial/engineering swarm intelligence system with numerous applications to real-world problems.

10
http://en.wikipedia.org/wiki/Parameter_spacehttp://en.wikipedia.org/wiki/Parameter_spacehttp://en.wikipedia.org/wiki/Pheromonehttp://en.wikipedia.org/wiki/Pheromonehttp://en.wikipedia.org/wiki/Swarm_intelligence#cite_note-3http://en.wikipedia.org/wiki/Bees_algorithmhttp://en.wikipedia.org/wiki/Foraginghttp://en.wikipedia.org/wiki/Honey_beehttp://www.scholarpedia.org/article/Graph_Theoryhttp://www.scholarpedia.org/article/Graph_Theoryhttp://www.scholarpedia.org/article/Network_sciencehttp://en.wikipedia.org/wiki/Parameter_spacehttp://en.wikipedia.org/wiki/Pheromonehttp://en.wikipedia.org/wiki/Swarm_intelligence#cite_note-3http://en.wikipedia.org/wiki/Bees_algorithmhttp://en.wikipedia.org/wiki/Foraginghttp://en.wikipedia.org/wiki/Honey_beehttp://www.scholarpedia.org/article/Graph_Theoryhttp://www.scholarpedia.org/article/Network_science


11/30

5.2 Particle Swarm Optimization

Particle swarm optimization (PSO) is a global optimization algorithm for dealing with problems

in which a best solution can be represented as a point or surface in an n-dimensional space. Hypotheses

are plotted in this space and seeded with an initial velocity, as well as a communication channel between

the particles. Particles then move through the solution space, and are evaluated according to

some fitness criterion after each timestep. Over time, particles are accelerated towards those particles

within their communication grouping which have better fitness values. The main advantage of such an

approach over other global minimization strategies such as simulated annealing is that the large number

of members that make up the particle swarm make the technique impressively resilient to the problem

of local minima .

Particle swarm optimization (Kennedy and Eberhart 1995; Kennedy, Eberhart and Shi, 2001) is a

population based stochastic optimization technique for the solution of continuous optimization problems. It is

inspired by social behaviors in flocks of birds and schools of fish. In particle swarm optimization (PSO), a set of

software agents called particles search for good solutions to a given continuous optimization problem. Each

particle is a solution of the considered problem and uses its own experience and the experience of neighbor

particles to choose how to move in the search space. In practice, in the initialization phase each particle is given a

random initial position and an initial velocity. The position of the particle represents a solution of the problem

and has therefore a value, given by the objective function. While moving in the search space, particles memorize

the position of the best solution they found. At each iteration of the algorithm, each particle moves with a velocity

that is a weighted sum of three components: the old velocity, a velocity component that drives the particle towards11
http://en.wikipedia.org/wiki/Fitness_(biology)http://en.wikipedia.org/wiki/Fitness_(biology)http://en.wikipedia.org/wiki/Fitness_(biology)


12/30

the location in the search space where it previously found the best solution so far, and a velocity component that

drives the particle towards the location in the search space where the neighbor particles found the best solution so

far. PSO has been applied to many different problems and is another example of successful artificial/engineering

swarm intelligence system.

6. REAL WORLD INSECT EXAMPLES

6.1 Bees12


13/30

Colony cooperation

Regulate hive temperature

Efficiency via Specialization: division of labour in the colony

13


14/30


15/30

Pulp foragers, water foragers & builders

Complex nests

Horizontal columns

Protective covering

Central entrance hole

6.3 Termites

15


16/30

Cone-shaped outer walls and ventilation ducts

Brood chambers in central hive

Spiral cooling vents

Support pillars

6.4 Ants

16


17/30

Organizing highways to and from their foraging sites by leaving pheromone trails

Form chains from their own bodies to create a bridge to pull and hold leafs together with silk

Division of labour between major and minor ants

6.5 Summary of the insects

The complexity and sophistication of 17


18/30


19/30


20/30

Light intensity is associated with attractiveness of a firefly, and such attraction enable the fireflies with the ability to

subdivide into small groups and each subgroup swarm around the local modes. Therefore, firefly algorithm is

specially suitable for multimodal optimization problems. In fact, FA has been applied in continuous

optimization, traveling salesman problem, clustering, image processing and feature selection.

Gravitational search algorithm

Gravitational search algorithm (GSA) is constructed based on the law of gravity and the notion of mass

interactions. The GSA algorithm uses the theory of Newtonian physics and its searcher agents are the collection

of masses. In GSA, there is an isolated system of masses. Using the gravitational force, every mass in the system

can see the situation of other masses. The gravitational force is therefore a way of transferring information

between different masses (Rashedi, Nezamabadi-pour and Saryazdi 2009). In GSA, agents are considered asobjects and their performance is measured by their masses. All these objects attract each other by a gravity force,

and this force causes a movement of all objects globally towards the objects with heavier masses. The heavy

masses correspond to good solutions of the problem. The position of the agent corresponds to a solution of the

problem, and its mass is determined using a fitness function. By lapse of time, masses are attracted by the

heaviest mass. We hope that this mass would present an optimum solution in the search space. The GSA could

be considered as an isolated system of masses. It is like a small artificial world of masses obeying the Newtonian

laws of gravitation and motion (Rashedi, Nezamabadi-pour and Saryazdi 2009). A multi-objective variant of GSA,

called Non-dominated Sorting Gravitational Search Algorithm (NSGSA), was proposed by Nobahari and

Nikusokhan in 2011.

Intelligent water drops

Intelligent Water Drops algorithm (IWD) is a swarm-based nature-inspired optimization algorithm, which has

been inspired by natural rivers and how they find almost optimal paths to their destination. These near optimal or

optimal paths follow from actions and reactions occurring among the water drops and the water drops with their

riverbeds. In the IWD algorithm, several artificial water drops cooperate to change their environment in such a

way that the optimal path is revealed as the one with the lowest soil on its links. The solutions are incrementally

constructed by the IWD algorithm. Consequently, the IWD algorithm is generally a constructive population-based

optimization algorithm.

20


21/30

Multi-swarm optimization

Multi-swarm optimization is a variant of particle swarm optimization (PSO) based on the use of multiple sub

swarms instead of one (standard) swarm. The general approach in multi-swarm optimization is that each sub

swarm focuses on a specific region while a specific diversification method decides where and when to launch the

sub-swarms. The multi-swarm framework is especially fitted for the optimization on multi-modal problems, where

multiple (local) optima exist.

Particle swarm optimization

Particle swarm optimization (PSO) is a global optimization algorithm for dealing with problems in which a best

solution can be represented as a point or surface in an n-dimensional space. Hypotheses are plotted in this space

and seeded with an initial velocity, as well as a communication channel between the particles. Particles thenmove through the solution space, and are evaluated according to some fitness criterion after each timestep. Over

time, particles are accelerated towards those particles within their communication grouping which have better

fitness values. The main advantage of such an approach over other global minimization strategies such

as simulated annealing is that the large number of members that make up the particle swarm make the technique

impressively resilient to the problem of local minima.

River formation dynamics

River formation dynamics (RFD) is an heuristic method similar to ant colony optimization (ACO). In fact, RFD can

be seen as a gradient version of ACO, based on copying how water forms rivers by eroding the ground and

depositing sediments. As water transforms the environment, altitudes of places are dynamically modified, and

decreasing gradients are constructed. The gradients are followed by subsequent drops to create new gradients,

reinforcing the best ones. By doing so, good solutions are given in the form of decreasing altitudes. This method

has been applied to solve different NP-complete problems (for example, the problems of finding a minimum

distances tree and finding a minimum spanning tree in a variable-cost graph). The gradient orientation of RFD

makes it specially suitable for solving these problems and provides a good tradeoff between finding good results

and not spending much computational time. In fact, RFD fits particularly well for problems consisting in forming a

kind of covering tree.

21


22/30

Self-propelled particles

Self-propelled particles (SPP), also referred to as the CouzinVicsek algorithm , was introduced in 1995 by Vicsek

and Couzin et al. as a special case of the boids model introduced in 1986 by Reynolds. A swarm is modelled in

SPP by a collection of particles that move with a constant speed but respond to a random perturbation by

adopting at each time increment the average direction of motion of the other particles in their local

neighbourhood. SPP models predict that swarming animals share certain properties at the group level, regardless

of the type of animals in the swarm. Swarming systems give rise to emergent behaviors which occur at many

different scales, some of which are turning out to be both universal and robust. It has become a challenge in

theoretical physics to find minimal statistical models that capture these behaviours.

Stochastic diffusion search

Stochastic diffusion search (SDS) is an agent-based probabilistic global search and optimization technique best

suited to problems where the objective function can be decomposed into multiple independent partial-functions.

Each agent maintains a hypothesis which is iteratively tested by evaluating a randomly selected partial objective

function parameterised by the agent's current hypothesis. In the standard version of SDS such partial function

evaluations are binary, resulting in each agent becoming active or inactive. Information on hypotheses is diffused

across the population via inter-agent communication. Unlike the stigmergic communication used in ACO, in SDS

agents communicate hypotheses via a one-to-one communication strategy analogous to the tandem running

procedure observed in some species of ant. A positive feedback mechanism ensures that, over time, a population

of agents stabilise around the global-best solution. SDS is both an efficient and robust search and optimization

algorithm, which has been extensively mathematically described.

Applications

Swarm Intelligence-based techniques can be used in a number of applications. The U.S. military is investigating

swarm techniques for controlling unmanned vehicles. The European Space Agency is thinking about an orbitalswarm for self assembly and interferometry. NASA is investigating the use of swarm technology for planetary

mapping. A 1992 paper by M. Anthony Lewis and George A. Bekey discusses the possibility of using swarm

intelligence to control nanobots within the body for the purpose of killing cancer tumors. Swarm intelligence has

also been applied for data mining.

22
http://en.wikipedia.org/wiki/Boidshttp://en.wikipedia.org/wiki/Boids


23/30

Crowd simulation

Artists are using swarm technology as a means of creating complex interactive systems or simulating crowds.

Stanley and Stella in: Breaking the Ice was the first movie to make use of swarm technology for rendering,

realistically depicting the movements of groups of fish and birds using the Boids system. Tim Burton's Batman

Returns also made use of swarm technology for showing the movements of a group of bats. The Lord of the

Rings film trilogy made use of similar technology, known as Massive, during battle scenes. Swarm technology is

particularly attractive because it is cheap, robust, and simple.

Airlines have used swarm theory to simulate passengers boarding a plane. Southwest Airlines

researcher Douglas A. Lawson used an ant-based computer simulation employing only six interaction rules to

evaluate boarding times using various boarding methods.(Miller, 2010, xii-xviii).

Ant-based routing

The use of Swarm Intelligence in Telecommunication Networks has also been researched, in the form of Ant

Based Routing. This was pioneered separately by Dorigo et al. and Hewlett Packard in the mid-1990s, with a

number of variations since. Basically this uses a probabilistic routing table rewarding/reinforcing the route

successfully traversed by each "ant" (a small control packet) which flood the network. Reinforcement of the route

in the forwards, reverse direction and both simultaneously have been researched: backwards reinforcement

requires a symmetric network and couples the two directions together; forwards reinforcement rewards a route

before the outcome is known (but then you pay for the cinema before you know how good the film is). As the

system behaves stochastically and is therefore lacking repeatability, there are large hurdles to commercial

deployment. Mobile media and new technologies have the potential to change the threshold for collective action

due to swarm intelligence (Rheingold: 2002, P175).

Airlines have also used ant-based routing in assigning aircraft arrivals to airport gates. At Southwest Airlines a

software program uses swarm theory, or swarm intelligencethe idea that a colony of ants works better than one

alone. Each pilot acts like an ant searching for the best airport gate. "The pilot learns from his experience what's

the best for him, and it turns out that that's the best solution for the airline," Douglas A. Lawson explains. As

aresult, the "colony" of pilots always go to gates they can arrive at and depart from quickly. The program can even

alert a pilot of plane back-ups before they happen. "We can anticipate that it's going to happen, so we'll have a

gate available," Lawson says.

23
http://en.wikipedia.org/wiki/Boidshttp://en.wikipedia.org/wiki/Boidshttp://en.wikipedia.org/wiki/Boids


24/30

7. SWARM INTELLIGENCE IN THEORY

7.1 An In-depth Look at Real Ant Behaviour

24


25/30

7.2 Interrupt The Flow

7.3 The Path Thickens!

25


26/30

7.4 The New Shortest Path

7.5 Adapting to Environment Changes

26


27/30

8. WELCOME TO THE REAL WORLD

27


28/30

8.1 Robots

28


29/30

Collective task completion

No need for overly complex algorithms

Adaptable to changing environment

8.2 Robot feeding demo

Routing packets to destination in shortest time

Similar to Shortest Route

Statistics kept from prior routing (learning from experience)

8.3 Swarm Robotics

Most important application area of Swarm Intelligence

Swarms provide the possibility of enhanced task performance, high reliability (fault tolerance),

low unit complexity and decreased cost over traditional robotic systems

Can accomplish some tasks that would be impossible for a single robot to achieve.

Swarm robots can be applied to many fields, such as flexible manufacturing systems, spacecraft,

inspection/maintenance, construction, agriculture, and medicine work

9. APPLICATIONS

29


30/30

Massive (Multiple Agent Simulation System in Virtual Environment) Software.

Developed Stephen Regelous for visual effects industry.

Snowbots

Developed Sandia National laboratory .

swarm intelligencce

Documents