Continuous approximation models for some modern logistical problems

Continuous approximation modelling is a logistical paradigm with origins at UC Berkeley in which detailed operational data is replaced by statistical summaries to analyze large-scale routing and distribution problems. This presentation introduces some new applications of computational geometry and geometric probability theory to study modern problems in last-mile delivery, districting, and order fulfillment. Building on the seminal 1959 Beardwood-Halton-Hammersley theorem, we give analytical frameworks that characterize the optimal solutions for various problems in an asymptotic limit as demand becomes large.

Speaker: John Gunnar Carlsson, University of Southern California

Room 3108