×

DISTRIBUTED STOCHASTIC CLUSTERING FOR AUTOMATED FORMATION OF CONNECTED NETWORKS OF AGENTS

  • US 20130179487A1
  • Filed: 01/10/2012
  • Published: 07/11/2013
  • Est. Priority Date: 01/10/2012
  • Status: Active Grant
First Claim
Patent Images

1. A method for forming a hierarchical network from an arbitrary set of agents, comprising steps for:

  • initially designating each agent of an arbitrary set of agents as a “

    head node”

    of its own hierarchical cluster;

    initiating an iterative process comprising steps for;

    causing each head node of each hierarchical cluster to randomly choose between a role as a “

    requestor” and

    a role as a “

    listener”

    ;

    for each requestor, randomly selecting an agent from the set of agents and sending a merge request to that agent;

    for each agent receiving a merge request, recursively passing that merge request to that agents head node;

    for each head node acting as a listener that receives only a single merge request, causing the head node to accept the merge request;

    for each head node acting as a listener that receives multiple merge requests, causing the head node to select a single one of the multiple merge requests;

    for all accepted merge requests, performing a separate disjoint merge to merge the hierarchical cluster of the requestor as a child cluster to the agent receiving the corresponding merge request; and

    repeating the steps of the iterative process until only a single hierarchical cluster remains for the arbitrary set of agents.

View all claims
  • 2 Assignments
Timeline View
Assignment View
    ×
    ×