Supplementary Material for "Making Better Design Decisions in Swarm Engineering by Aligning Intuition with Analysis"

Overview

This page provides supplementary material for the AAS paper. The foraging algorithms shown on this page are:

  • CRW
  • DPO
  • STOCHM
  • STOCHX

See AAMAS 2020,IJCAI 2019 for technical explanations of the algorithms.

All videos were recorded/generated using SIERRA. All videos are 5,000 seconds of simulated time long.

Blocks to be collected are represented as black squares, and the nest is a light gray with yellow lights above it which robots use for phototaxis. Caches are a dark gray (STOCHM, STOCHX only).

ARGoS Visual Videos

Scenario 1 (small warehouse), 32x16 arena with 128 robots:

  • CRW (left), DPO (right)

  • STOCHM (left), STOCHX (right)

Diagnostic Videos

Scenario 1 (small warehouse), 32x16 arena with single source block distribution, 128 robots:

  • CRW: spatial searching distribution (left), spatial interference distribution (right):

  • DPO: spatial searching distribution (left), spatial interference distribution (right):

  • STOCHM: spatial searching distribution (left), spatial interference distribution (right):

  • STOCHX: spatial searching distribution (left), spatial interference distribution (right):

Scenario 1 (medium warehouse), 48x48 arena with random block distribution, 256 robots:

  • CRW: spatial block pickup distribution (left), spatial occupancy distribution (right):

  • DPO: spatial block pickup distribution (left), spatial occupancy distribution (right):

  • STOCHM: spatial block pickup distribution (left), spatial occupancy distribution (right):

  • STOCHX: spatial block pickup distribution (left), spatial occupancy distribution (right):

John Harwell
John Harwell
Researcher and Engineer

Experienced embedded systems engineer whose research interests include multi-agent modeling and behaviors, swarm intelligence, bio-inspired algorithms and multi-robot systems, and computational optimization.