+ All Categories
Home > Documents > Sworm Intelligence

Sworm Intelligence

Date post: 03-Apr-2018
Category:
Upload: faizan-hyder
View: 218 times
Download: 0 times
Share this document with a friend

of 22

Transcript
  • 7/28/2019 Sworm Intelligence

    1/22

  • 7/28/2019 Sworm Intelligence

    2/22

    Name: Faizan HyderRoll No: 08040619-098

  • 7/28/2019 Sworm Intelligence

    3/22

    Swarm Intelligence

    What is Swarm?A large or dense group of insects.

    Swarm intelligence (SI) is the collective behaviorof decentralized, self-organized systems, natural orartificial.

  • 7/28/2019 Sworm Intelligence

    4/22

    Characteristics of Swarms

    Composed of many individuals Individuals are Of the same kind

    Local interaction based on simple

    rules Self-organization

  • 7/28/2019 Sworm Intelligence

    5/22

    Two principles of Swarm Intelligence

    Self-organization is based on: activity amplification by positive feedback

    activity balancing by negative feedback

    multiple interactions

    Stigmergy - stimulation by work - is basedon:

    Stigmergy is a mechanism of indirect coordinationbetween agents or actions

    Work as behavioural response to the environmental

    state work that does not depend on specific agents

  • 7/28/2019 Sworm Intelligence

    6/22

    Ant Colony Optimization:

    A heuristic optimization method forshortest path and other optimizationproblems which borrows ideas frombiological ants.

    ACO is implemented as a team ofintelligent agents which simulate theants behavior, walking around the graph

    representing the problem to solve usingmechanisms of cooperation andadaptation.

  • 7/28/2019 Sworm Intelligence

    7/22

    Ant Colony Optimization - BiologicalInspiration

    Inspired by foraging behavior of ants. Ants find shortest path to food source

    from nest.

    Ants deposit pheromone along traveledpath which is used by other ants to followthe trail.

    This kind of indirect communication via the

    local environment is called stigmergy.

  • 7/28/2019 Sworm Intelligence

    8/22

    Pheromone

    Substances secreted by an animal toinfluence the behavior of other animalsof the same species.

    Two minutes after a ant has marked a

    trail, the pheromone has evaporated toa point that is below detection level.

    Only if food is found does the workeremit a pheromone trail and only on theway back to the nest.

  • 7/28/2019 Sworm Intelligence

    9/22

    Pheromone(Cont..)

    If a worker loses the trail, it travels insmall circles until the trail is eitherfound, or another worker interceptsthe lost ant by antennae touching

    and guides it back to the trail

  • 7/28/2019 Sworm Intelligence

    10/22

    Foraging behavior of Ants

    2 ants start with equal probability ofgoing on either path.

  • 7/28/2019 Sworm Intelligence

    11/22

    Foraging behavior of Ants

    The ant on shorter path has a shorterto-and-fro time from its nest to thefood.

  • 7/28/2019 Sworm Intelligence

    12/22

    Foraging behavior of Ants

    The density of pheromone on theshorter path is higher because of 2passes by the ant (as compared to 1by the other).

  • 7/28/2019 Sworm Intelligence

    13/22

    Foraging behavior of Ants

    The next ant takes the shorter route.

  • 7/28/2019 Sworm Intelligence

    14/22

    Foraging behavior of Ants

    Over many iterations, more ants beginusing the path with higher pheromone,thereby further reinforcing it.

  • 7/28/2019 Sworm Intelligence

    15/22

    Foraging behavior of Ants

    After some time, the shorter path isalmost exclusively used.

  • 7/28/2019 Sworm Intelligence

    16/22

    ACO algorithm

    Main steps of the ACO algorithm aregiven below:

    Pheromone trail initialization

    Solution construction using pheromone

    trail Each ant constructs a complete solution to

    the problem according to a probabilistic

    State transition rule. The state transition

    rule depends mainly on the state of thepheromone .

    Pheromone trailupdate.

  • 7/28/2019 Sworm Intelligence

    17/22

    Ant Colony Optimization Algorithm

  • 7/28/2019 Sworm Intelligence

    18/22

    Choosing Next Node

    When being at a node, an ant chooses to go toa unvisited node at time twith a probabilitygiven by

    wherejkiis, the setof cities which antkhas not yet visited; i,j(t) is the pheromone value on the edge(i, j) at the time t, is

    the weight of pheromone; i,j (t) is a priori available heuristicinformation on the edge (i, j) at the time t, is the weight ofheuristic information. Two parameters and determine therelative influence of pheromone trail and heuristic information.

  • 7/28/2019 Sworm Intelligence

    19/22

    Pheromone Update

    When all the ants have completed asolution, the trails are updated by

    where is the pheromone trailevaporation rate (0

  • 7/28/2019 Sworm Intelligence

    20/22

    Advantages

    For TSPs (Traveling Salesman Problem), relatively

    efficient

    Performs better against other global optimizationtechniques for TSP (neural net, geneticalgorithms)

    Compared to GAs (Genetic Algorithms):lessaffected by poor initial solutions (due tocombination of random path selection and colonymemory)

    Can be used in dynamic applications (adapts tochanges such as new distances, etc.)

  • 7/28/2019 Sworm Intelligence

    21/22

  • 7/28/2019 Sworm Intelligence

    22/22


Recommended