Geometric Optimization and Combinatorial Homological Programming
Abstract
New mathematical methods for distributed optimization and related problems were generated using concepts from algebraic topology and sheaf theory. In particular, the Hodge Laplacian was extended to sheaves of data in the context of distributed optimization, indoor mapping, network communications, and learning. The results greatly extend known results from spectral graph theory and simultaneous localization and mapping.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 03, 2019
- Accession Number
- AD1090239
Entities
People
- Robert Ghrist
Organizations
- University of Pennsylvania