A dynamic-programming-styled algorithm for a class of multi-agent optimal task assignment

Guang Yang, Vikram Kapila, Ravi Vaidyanathan

Research output: Contribution to conferencePaper

Abstract

In this paper, we use a dynamic programming formulation to address a class of multi-agent task assignment problems that arise in the study of fuel optimal control of multiple agents. The fuel optimal multi-agent control is highly relevant to multiple spacecraft formation reconfiguration, an area of intense current research activity. Based on the recurrence relation derived from the celebrated principle of optimality, we develop an algorithm with a distributed computational architecture for the global optimal task assignment. In addition, we propose a communication protocol to facilitate decentralized decision making among agents. Illustrative studies are included to demonstrate the efficacy of the proposed multi-agent optimal task assignment algorithm.

Original languageEnglish (US)
Pages293-299
Number of pages7
StatePublished - 2002
Event2001 ASME International Mechanical Engineering Congress and Exposition - New York, NY, United States
Duration: Nov 11 2001Nov 16 2001

Conference

Conference2001 ASME International Mechanical Engineering Congress and Exposition
CountryUnited States
CityNew York, NY
Period11/11/0111/16/01

ASJC Scopus subject areas

  • Mechanical Engineering
  • Software

Fingerprint Dive into the research topics of 'A dynamic-programming-styled algorithm for a class of multi-agent optimal task assignment'. Together they form a unique fingerprint.

  • Cite this

    Yang, G., Kapila, V., & Vaidyanathan, R. (2002). A dynamic-programming-styled algorithm for a class of multi-agent optimal task assignment. 293-299. Paper presented at 2001 ASME International Mechanical Engineering Congress and Exposition, New York, NY, United States.