Loading...

Improved algorithms for distributed balanced clustering

Mirjalali, K ; Sharif University of Technology | 2020

306 Viewed
  1. Type of Document: Article
  2. DOI: 10.1007/978-3-030-57852-7_6
  3. Publisher: Springer Science and Business Media Deutschland GmbH , 2020
  4. Abstract:
  5. We study a weighted balanced version of the k-center problem, where each center has a fixed capacity, and each element has an arbitrary demand. The objective is to assign demands of the elements to the centers, so as the total demand assigned to each center does not exceed its capacity, while the maximum distance between centers and their assigned elements is minimized. We present a deterministic O(1)-approximation algorithm for this generalized version of the k-center problem in the distributed setting, where data is partitioned among a number of machines. Our algorithm substantially improves the approximation factor of the current best randomized algorithm available for the problem. We also show that the approximation factor of our algorithm can be improved to, when the underlying metric space has a bounded doubling dimension. © 2020, IFIP International Federation for Information Processing
  6. Keywords:
  7. Approximation algorithms ; Approximation factor ; Balanced clustering ; Demand assigned ; Doubling dimensions ; K-center problem ; Maximum distance ; Metric spaces ; Randomized Algorithms ; Clustering algorithms
  8. Source: 3rd IFIP WG 1.8 International Conference on Topics in Theoretical Computer Science, TTCS 2020, 1 July 2020 through 2 July 2020 ; Volume 12281 LNCS , 2020 , Pages 72-84
  9. URL: https://www.springer.com/gp/book/9783030217174