+ All Categories
Home > Documents > Swarm Intelligenge_gaurav Agrawal

Swarm Intelligenge_gaurav Agrawal

Date post: 07-Apr-2018
Category:
Upload: gaurav-agrawal
View: 218 times
Download: 0 times
Share this document with a friend

of 20

Transcript
  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    1/20

    Swarm Intelligence

    Gaurav Agrawal

    Phd scholar (computer science)

    Sharda University

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    2/20

    OUTLINE

    Background

    What is a Swarm Intelligence (SI)?

    Examples from nature

    Origins and Inspirations of SI

    Ant Colony Optimization

    Particle Swarm Optimization

    Properties of a Swarm Intelligence System

    ADVANTAGES OF SI

    SI APPLICATIONS

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    3/20

    swarm behavior

    The behaviors of a flock of

    birds, a group of ants, a school

    of fish, etc.,were the field of

    study during earlier days. Such

    collective motion of insectsand birds is known to as

    swarm behavior.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    4/20

    WHAT IS A SWARM?

    A loosely structured collection of interacting agents

    Agents:

    Individuals that belong to a group (but are not necessarily

    identical)

    They contribute to and benefit from the group

    They can recognize, communicate, and/or interact with each

    other

    A swarm is better understood if thought of as agents exhibiting a

    collective behavior

    A type of self-organization emerges from the collection of actions of thegroup.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    5/20

    SWARM INTELLIGENCE (SI)

    Gerardo Beni and Jing Wang introduced the term swarm intelligence

    in a 1989.

    Swarm intelligence is the collective intelligence of groups of simple

    autonomous agents.

    The autonomous agent is a subsystem that interacts

    with its environment, relatively independently from all other agents.

    There is no global plan or leader to control the entire group of

    autonomous agents.

    Based on group behavior found in nature

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    6/20

    Swarm intelligence techniques are population

    based stochastic methods used in

    combinatorial optimization problems

    in which the collective behavior of relatively

    simple individuals arises from their local

    interactions with their environment to produce

    functional global patterns.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    7/20

    EXAMPLES OF SWARMS IN

    NATURE:

    Classic Example: Swarm ofBees

    Can be extended to other similar systems:

    Ant colony

    Agents: ants

    Flock of birds

    Agents: birds

    Traffic

    Agents: cars

    CrowdAgents: humans

    Immune system

    Agents: cells and molecules

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    8/20

    Boid model

    In 1987, a boid = Bird-oids (bird like) model was created by Reynold.

    This boid is a distributed behavioral model, which is used to simulate

    the motion of a flock of birds on a personal computer.

    Each boid serves as an independent actor that navigates based on

    its own perception of the dynamic environment.

    There are a certain set of rules that are to be observed by the boid.

    The avoidance rule states that an individual boid must move awayfrom boids that are too close, so as to reduce the chance of in-air

    collisions.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    9/20

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    10/20

    TWO COMMON SI ALGORITHMS

    Ant Colony Optimization

    Particle Swarm Optimization

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    11/20

    PARTICLE SWARMOPTIMIZATION (PSO)

    Particle Swarm Optimization (PSO) was proposed by JamesKennedy and R. C. Eberhart in 1995, inspired by social behavior oforganisms such as bird flocking and fish schooling.

    The underlying concept is that, for every time instant, the velocity ofeach particle also known as the potential solution, changes betweenits pbest and lbest locations.

    The particle associated with the best solution (fitness value) seemsto be the leader and each particle keeps track of its coordinates inthe problem space.

    This fitness value is stored which is referred to as pbest. Another best value that is tracked by the particle swarm optimizer

    is the best value, obtained so far by any particle in the neighbors ofthe particle. This location is called lbest.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    12/20

    when a particle takes all the population as its topological neighbors,

    the best value is a global best and is called gbest.

    Thus a PSO system combines local search methods

    with global search methods

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    13/20

    Ant ColonyOptimization (ACO)

    Is a class of algorithms, whose firstmember, called Ant System, was initiallyproposed by Colorni, Dorigo and Maniezzo1992.

    In nature, ants usually wander randomly,and upon finding food return to their nestwhile laying down pheromone trails. The

    other ants find the path (pheromone trail),and follow the trail, returning and reinforcingit if they eventually find food. Thepheromone starts to evaporate as timepasses. If the time taken for an ant to traveldown the path and back again to the nest,the pheromone evaporates thereby making

    the path less prominent. A shorter path, incomparison will be visited by more ants (canbe described as a loop of positive feedback)and thus the pheromone density remainshigh for a longer time.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    14/20

    An ant runs more or less at random

    around the colony;

    1. On discovering food, it returns more

    or less directly to the nest, leavingin its path a trail of pheromone;

    2.These pheromones are attractive,

    nearby ants will be inclined to

    follow, more or less directly, the

    track

    3.Returning to the colony, these antswill strengthen the route;

    4. If there are two routes to reach the

    same food source then, in a given

    amount of time, the shorter one will

    be traveled by more ants than the

    long route;The short route will be increasingly

    enhanced, and therefore become

    more attractive;

    The long route will eventually isappear

    because pheromones are volatile;

    Eventually, all the ants havedetermined and therefore "chosen"

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    15/20

    ACO is implemented as a collective group of intelligent agents,which simulate the ants behavior, walking around the graphrepresenting the problem to solve using mechanisms of cooperationand adaptation. ACO algorithm requires the following definitions:

    The problem needs to be represented appropriately, which would

    allow the ants to incrementally update the solutions through the useof a probabilistic transition rules, based on the amount ofpheromone in the trail and other problem specific knowledge. It isalso important to enforce a strategy to construct only valid solutions

    corresponding to the problem definition

    A problem-dependent heuristic function that measures the qualityof components that can be added to the current partial solution

    A rule set for pheromone updating, which specifies how to modify

    the pheromone value .

    A probabilistic transition rule based on the value of the heuristicfunction and the pheromone value that is used to iteratively

    construct a solution

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    16/20

    Properties of a Swarm Intelligence System

    Unity:A swarm is a combination of several individuals.

    Fault tolerance: Swarm intelligent processes do not rely on acentralized control mechanism. Therefore the loss of a few nodes orlinks does not result in catastrophic failure.

    Rule-based behavior:A certain set of rules are observed by the

    individuals that exploit only local information that the individualsexchange directly or through the environment.

    Autonomy: The overall behavior of the swarm system is selforganized, it does not depend on orders external to the system

    itself. No human supervision is required.

    Scalability: Population of the agents can be adapted according to

    the network size. Scalability is also promoted by local anddistributed agent interactions.

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    17/20

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    18/20

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    19/20

    SI APPLICATIONS

    Military is applying SI techniques to control of unmanned

    vehicles

    NASA is applying SI techniques for planetary mapping

    Medical Research is trying SI based controls fornanobots to fight cancer (Swarm Robotics)

    SI techniques are applied to load balancing/ routing in

    telecommunication networks

    Entertainment industry is applying SI techniques forbattle and crowd scenes

  • 8/6/2019 Swarm Intelligenge_gaurav Agrawal

    20/20

    Thank You


Recommended