Conference 2021
Top image

Program LNMB conference
Invited Speakers
PhD student pitches
Return to LNMB Site

Natasja Sluijk (Eindhoven University of Technology) - The Chance-Constrained Two-Echelon Vehicle Routing Problem with Stochastic Demands
Supervisor: Tom van Woensel
Recorded full presentation

Abstract In this work, we present the first exact method for solving the two-echelon vehicle routing problem with stochastic demands. We formulate the two-echelon vehicle routing problem as a chance-constrained stochastic optimization problem and propose an efficient solution procedure based on column generation. Key to the efficiency of this procedure is the underlying labeling algorithm to generate new columns. To this extend we propose a novel labeling algorithm based on simultaneous labeling of second-echelon routes. To further enhance the performance of the solution procedure, we use statistical inference techniques to ensure that the probabilistic vehicle capacity constraints are met. Statistical bounds on the stochastic customer demands are imposed to reduce the number of customer combinations for which the capacity constraints need to be verified. The idea of statistical bounds can be applied to any set of options and type of stochasticity.