Download PDFOpen PDF in browser

Topology Planning in Swarm Production System: Framework and Optimization

EasyChair Preprint 9432

16 pagesDate: December 8, 2022

Abstract

Swarm Production System (SPS) aims to be an agile and resilient Reconfigurable manufacturing system (RMS) paradigm that incorporates mobile workstations and transport robots on the factory production floor. This paper primarily focuses on SPS's initial but recurring planning stage termed topology planning, which dynamically changes throughout the production runtime with spatially adaptive workstations and transporters handled exclusively by a Topology Manager (TM). TM is essential to mass-produce multiple variants with the optimal positioning of the workstations and allow a collision-free logical path to the product carrying transport robots. TM is a bridge to enable SPS to integrate with general planning and scheduling systems like ERP and MES and is comprised of a Topology Planner (TP) that evaluates the ideal configuration of on factory floor for a batch of product mix and a Reconfiguration Decision System (RDS) that decides on applying the estimated new configuration during the batch changeover. The paper proposes a framework for the TM to identify its essential functionalities, responsibilities and working principle in a swarm production system. The paper also describes a grid-based heuristic 2 stage approach applicable to two-dimensional spatial problems to reduce the complexity of an NP-hard problem. The first stage focuses on finding an initial placement with a Force-directed Graph-theory, and the second optimization stage with a heuristically developed spanning tree reducing the overlapping edges. A stochastic mathematical scheduling model evaluates the performance of two different topologies and why a spanning-tree topology offers a better trade-off in terms of throughput.

Keyphrases: Reconfigurable Manufacturing System, Reconfiguration Decision System, Swarm Production, Topology Manager, force-directed layout, graph theory, spanning tree, stochastic model

BibTeX entry
BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:
@booklet{EasyChair:9432,
  author    = {Akshay Avhad and Casper Schou and Ole Madsen},
  title     = {Topology Planning in Swarm Production System: Framework and Optimization},
  howpublished = {EasyChair Preprint 9432},
  year      = {EasyChair, 2022}}
Download PDFOpen PDF in browser