Bi-objective facility location under uncertainty with an application in last-mile disaster relief

Area of Science:

• Operations Research
• Disaster Management
• Optimization

Background:

• Real-world problems often involve multiple, conflicting objectives and data uncertainty.
• Decision-makers require understanding of objective trade-offs under various uncertainty levels.
• Designing effective last-mile networks for disaster relief presents significant challenges.

Purpose of the Study:

• To develop and analyze a two-stage bi-objective capacitated model for last-mile disaster relief network design.
• To incorporate demand uncertainty into the network design problem.
• To evaluate different optimization approaches for determining Pareto frontiers.

Main Methods:

• Scenario-based two-stage risk-neutral stochastic programming.
• Adaptive (two-stage) robust optimization.
• Two-stage risk-averse stochastic approach using conditional value-at-risk (CVaR).
• Criterion space search frameworks ($$-constraint and balanced box methods).
• Matheuristic technique for large-scale instances.

Main Results:

• The study successfully applies various optimization techniques to a bi-objective last-mile network design problem.
• Different approaches provide insights into managing demand uncertainty and objective trade-offs.
• The matheuristic effectively generates high-quality Pareto frontier approximations for large problems.

Conclusions:

• The proposed models and methods offer valuable tools for designing resilient last-mile disaster relief networks.
• Understanding the Pareto frontier is crucial for informed decision-making in uncertain environments.
• The research provides a robust framework applicable to real-world disaster scenarios, demonstrated by a case study in Senegal.