PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
June 1, 1982IEEE Transactions on Automatic Control222 citations

Distributed dynamic programming

View Full Paper
DBDimitri P. Bertsekas

Key Points

  • To develop distributed algorithms for solving dynamic programming problems with weak computational assumptions.
  • Developed a model of asynchronous distributed computation with minimal assumptions.
  • Enabled multiple processors to compute simultaneously while exchanging information.
  • Applied the model to a broad class of problems, including shortest path and stochastic optimal control problems.
  • Algorithm reduces to a routing method for ARPANET when applied to shortest path problems.
  • Demonstrates effective coordination among processors with weak assumptions on timing and information needs.
  • Broad applicability to various dynamic programming issues across different contexts.

Abstract

We consider distributed algorithms for solving dynamic programming problems whereby several processors participate simultaneously in the computation while maintaining coordination by information exchange via communication links. A model of asynchronous distributed computation is developed which requires very weak assumptions on the ordering of computations, the timing of information exchange, the amount of local information needed at each computation node, and the initial conditions for the algorithm. The class of problems considered is very broad and includes shortest path problems, and finite and infinite horizon stochastic optimal control problems. When specialized to a shortest path problem the algorithm reduces to the algorithm originally implemented for routing of messages in the ARPANET.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Dimitri P. Bertsekas (1982) studied this question.

synapsesocial.com/papers/6a11e7149ffe35dda08e1074https://doi.org/10.1109/tac.1982.1102980
Ask AI
Helpful
Bookmark
Share
View Full Paper