/dcop_simulator_5

Primary LanguagePython

Sync DCOP-MST Simulator (version 5)

Models

DCOP

Distributed Search.

DCOP_MST

Distributed Search with Mobile Sensor Teams.

Algorithms

  • Random
  • DSA_MST
  • CADSA
  • DSSA
  • Max-sum_MST
  • Max-sum_MST with breakdowns
  • CAMS

Big Experiments

  • 7 aforementioned algorithms
  • 200 steps to each problem
  • 20 problems per map per targets' type
  • 4 maps: empy, random, warehouse, room
  • 2 types of targets: static and dynamic
  • 20 small iterations for CAMS
  • 2 metrics: collision (col) metric and remained coverage requirement (rcr) metric

Agents:

  • number of agents is fixed for everything (in small maps - 20, in big maps - )
  • SR of agents is fixed for everything
  • MR of agents is fixed for everything
  • Cred of agents is fixed for everything

Targets:

  • number of targets is fixed for everything (in small maps - 10, in big maps - )
  • Req of targets is fixed for everything
  • Position of targets fixed while static
  • Position of targets changes every 40 steps while dynamic

For one type of target and one map there is the following data structure:

import numpy as np
# per type of targets per map: 
data_structure_for_json = { 
    'alg_name':  # (7 of them)
    np.zeros((200, 20))  # [[200(steps) x 20(problems) matrix]]
}
map Static Targets Dynamic Targets
empty (empty-48-48)
random (random-32-32-10)
warehouse (warehouse-10-20-10-2-1)
room (lt_gallowstemplar_n)

Credits