Optimal Transport for Concave Cost Functions

par Grigorii Buklei (GSSI L’Aquila)

mercredi 1 avr. 2026, 10:30 → 12:00 Europe/Paris

112 (Braconnier )

In a pizza delivery scenario, we have n points x_1, …, x_n (couriers) and y_1, …, y_n (customers) on a line, and we want to match them so that the total delivery cost is minimised. Motivated by economic considerations, we will consider the case where the cost of a matching is concave in the distance, say |x-y|^p for p \in (0,1). Different cost functions of this type can impose different optimal courier assignments. The goal of the talk is to discuss how to obtain quantitative bounds on this number. Time permitting, I will also touch on other OT-related projects I am working on.