ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 1905.07647
19
11

On greedy heuristics for computing D-efficient saturated subsets

18 May 2019
Radoslav Harman
Samuel Rosa
ArXivPDFHTML
Abstract

Let F\mathcal{F}F be a set consisting of nnn real vectors of dimension m≤nm \leq nm≤n. For any saturated, i.e., mmm-element, subset S\mathcal{S}S of F\mathcal{F}F, let vol(S)\mathrm{vol}(\mathcal{S})vol(S) be the volume of the parallelotope formed by the vectors of S\mathcal{S}S. A set S∗\mathcal{S}^*S∗ is called a DDD-optimal saturated subset of F\mathcal{F}F, if it maximizes vol(S)\mathrm{vol}(\mathcal{S})vol(S) among all saturated subsets of F\mathcal{F}F. In this paper, we propose two greedy heuristics for the construction of saturated subsets performing well with respect to the criterion of DDD-optimality: an improvement of the method suggested by Galil and Kiefer for the initiation of DDD-optimal experimental design algorithms, and a modification of the Kumar-Yildirim method, the original version of which was proposed for the initiation of the minimum-volume enclosing ellipsoid algorithms. We provide geometric and analytic insights into the two methods, and compare them to the commonly used random and regularized greedy heuristics. We also suggest variants of the greedy methods for a large set F\mathcal{F}F, for the construction of DDD-efficient non-saturated subsets, and for alternative optimality criteria.

View on arXiv
Comments on this paper