Abstract:Service network design (SND) problem is designing the service selection and flow distribution to satisfy demands under limited assets. A growing body of research has shown that ignoring uncertainty can increase total operational costs when fluctuations arise and must be managed in the SND. In this case, stochastic programming (SP) and robust optimization (RO) are commonly utilized. Nevertheless, SP requires precise distribution data that may not be accessible in some situations, while RO typically delivers an over-conservative solution under the worst case. In order to avoid the disadvantages, we focus on distributionally robust optimization (DRO) for the uncertain SND. DRO formulation does not need accurate distribution information of stochastic variables, which is useful for insufficient historical data or inaccurate distribution prediction situations. DRO utilizes restricted historical information to construct an ambiguity set for possible distributions rather than specifying an uncertain set artificially as RO does, whichÂ optimizes the formulation under the worst distribution and reduces the conservative degree. In this paper, we build three DRO formulations based on different ambiguity sets with varying trustable historical knowledge and reformulate them to be solvable as a mixed integer linear program, a second-order cone program, and a semi-definite program, respectively. A delayed constraint generation framework is designed to obtain optimal solutions for the formulations. We contrastively design several groups of experiments to evaluate the performance of distributionally robust SNDs and the difference among different methodologies, based on which we give some network design insights to managers.

Robust Optimization for Tree-Structured Stochastic Network Design

Robust Optimization of Hydrogen Network

Mixed-Integer Optimization for Bio-Inspired Robust Power Network Design

A Surrogate-Based Optimization Algorithm for Network Design Problems

Robust Optimization Using a Methodology Based on Cross Entropy Methods

Robust Optimization Model for A Dynamic Network Design Problem under Demand Uncertainty

Non-Probabilistic Robust Equilibrium Optimization of Complex Uncertain Structures

Robust Stochastic Optimization with Rare-Event Modeling

Robust Optimization Model of Bus Transit Network Design with Stochastic Travel Time

Network planning under demand uncertainty with robust optimization

Adjustable Robust Nonlinear Network Design under Demand Uncertainties

A Two-Stage Robust Optimization for Reliable Logistics Network Design via Evolutionary Computation

Robust Transportation Network Design under Demand Uncertainty

A Stochastic Benders Decomposition Scheme for Large-Scale Stochastic Network Design

Multistage Distributionally Robust Design of a Renewable Source Processing Network under Uncertainty

Network Flow Models for Robust Binary Optimization with Selective Adaptability

A two-stage stochastic programming model for transportation network protection

Robust optimization for water distribution systems least cost design

Distributionally Robust Optimization for Service Network Design Considering Different Trustable Moments of Historical Data

Measures and Optimization for Robustness and Vulnerability in Disconnected Networks