Selective solution pool pruning in multi-agent optimization systems for the transit route network design problem

Jeremy J. Blum, Scott E. Edsell

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Scopus citations

Abstract

Agent optimization systems have been used to find near optimal results for many intractable problems, including the Traveling Salesman problem and the Transit Route Network Design (TRND) problem. In the agent architectures used for these problems, agents collaborate by working with a common pool of solutions, creating, modifying, or destroying solutions in this pool. It has been posited that "adept destruction can compensate for inept construction" in these optimization systems [1]. This paper proposes a Selective Solution Pool Pruning agent that manages the diversification and intensification of the agent search process. This agent is instantiated for a system that is designed to optimize the TRND problem for a de facto benchmark transit network. The optimization system results show the Selective Solution Pool Pruning agent provides an effective means for managing the search process.

Original languageEnglish (US)
Title of host publication2011 14th International IEEE Conference on Intelligent Transportation Systems, ITSC 2011
Pages710-715
Number of pages6
DOIs
StatePublished - Dec 22 2011
Event14th IEEE International Intelligent Transportation Systems Conference, ITSC 2011 - Washington, DC, United States
Duration: Oct 5 2011Oct 7 2011

Publication series

NameIEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC

Other

Other14th IEEE International Intelligent Transportation Systems Conference, ITSC 2011
CountryUnited States
CityWashington, DC
Period10/5/1110/7/11

All Science Journal Classification (ASJC) codes

  • Automotive Engineering
  • Mechanical Engineering
  • Computer Science Applications

Fingerprint Dive into the research topics of 'Selective solution pool pruning in multi-agent optimization systems for the transit route network design problem'. Together they form a unique fingerprint.

Cite this