Joint Optimization of Fair Facility Allocation and Robust Inventory Management for Perishable Consumer Products

Saba Ghasemi Naraghi, Zheyu JiangSystems 2025 · DOI

Abstract

Perishable consumer products like food, cosmetics, and household chemicals face challenges in supply chain management due to limited shelf life and uncertainties in demand and transportation. To address some of these issues, this work proposes a robust optimization framework for jointly optimizing facility allocation and inventory management. The framework determines optimal locations for distribution centers and their assigned customers, as well as inventory policies that minimize the total costs related to transportation, distribution, and storage under uncertain demand in a robust setting. Specifically, we develop a two-stage mixed-integer linear programming (MILP) model that incorporates a First-In-First-Out (FIFO) inventory policy to reduce spoilage. The bilinear FIFO constraints are linearized to improve computational efficiency. Social equity is integrated by defining a fairness index and incorporating it in facility allocation. Demand uncertainty is tackled using a robust optimization approach with affine demand functions to handle multiple scenarios. The model is solved using row and column generation techniques for scalability. Overall, this robust optimization framework is expected to enhance supply chain resilience, reduce waste, and improve cost-effectiveness in managing perishable products.

Introduction

Managing supply chains for perishable products—such as fresh food, pharmaceuticals, and cosmetics—is a high-stakes balancing act. Overstocking leads to significant economic and environmental waste; in the United States alone, approximately 38% of food goes unsold. Conversely, understocking results in lost revenue, penalties, and reduced customer satisfaction.

Traditionally, facility location and inventory management have been treated as separate problems. However, these decisions are deeply interdependent. Furthermore, conventional models often overlook social equity, which can lead to disadvantaged communities having poor access to essential goods. This paper introduces a robust optimization framework that simultaneously addresses facility allocation and inventory management for perishable goods. By integrating a social fairness index and a linearized First-In-First-Out (FIFO) inventory policy into a two-stage robust model, the authors demonstrate how to balance cost-efficiency, waste reduction, and equitable community access under uncertain demand.

Methodology

The proposed framework is formulated as a two-stage Mixed-Integer Linear Programming (MILP) model designed to handle the complexities of perishable supply chains.

Two-Stage Decision Process

  1. First Stage (Strategic): This "here-and-now" stage involves decisions regarding the location and capacity of distribution centers (DCs) and their assigned customer regions. A critical feature is the inclusion of an $ε$-constraint for social fairness, which ensures that facility allocation meets a minimum threshold for serving areas characterized by high poverty and unemployment.
  2. Second Stage (Operational): This "wait-and-see" stage manages inventory levels across various demand scenarios. To minimize spoilage, the authors implement a FIFO policy. Since standard FIFO constraints are bilinear and non-convex, a linearization strategy is employed to maintain computational tractability.

Robust Optimization and Scalability

To account for the inherent uncertainty in demand, the model utilizes Robust Optimization with affine demand functions. Rather than optimizing for average demand, the framework prepares for the worst-case scenario within a defined uncertainty set. To ensure the model scales to large networks (such as a 50-state model), it is solved using row and column generation, an iterative technique that selectively adds the most challenging scenarios to the master problem.

Data and Implementation

The framework was evaluated using a hypothetical nationwide U.S. supply chain for four perishable food categories: dairy, meat, fruits, and vegetables.

  • Social Metrics: State-level social fairness indices were calculated using 2023-2024 data from the U.S. Census and Bureau of Labor Statistics, focusing on poverty and unemployment rates.
  • Demand Estimation: Average daily demand across all 50 states was estimated using data from the USDA and Statista.
  • Economic Parameters: Real-world estimates for land acquisition, construction, labor, energy, and transportation costs were used to parameterize the model.

Results and Discussion

The Fairness-Cost Trade-off

A key finding of the study is the quantifiable trade-off between social equity and system efficiency. By mapping a Pareto front, the researchers demonstrated that increasing the social fairness requirement ($ε$)—thereby placing facilities in more disadvantaged areas—leads to a continuous increase in total system costs. This provides policymakers with a clear "price tag" for social equity initiatives.

Pareto Front showing the trade-off between the social constraint (epsilon) and total system cost.

Spoilage Sensitivity and Robustness

The model reveals that spoilage is highly sensitive to the "spoilage cost factor" ($β^{spoil}$). If the penalty for spoilage is high enough (e.g., double the supply cost), the optimization naturally eliminates waste, even under demand uncertainty. However, there is a clear link between conservatism and waste: higher levels of robustness (preparing for more extreme demand fluctuations, denoted by $Γ$) generally lead to increased spoilage, highlighting the inherent cost of risk-aversion in perishable supply chains. In high-uncertainty scenarios with low spoilage penalties, the system recorded nearly 120 Mlbs of waste.

Analysis of spoilage levels across different demand uncertainty levels (Gamma) and spoilage cost factors.

Inventory "Weight Shift"

The researchers observed a "weight shift" phenomenon: at lower fairness levels where fewer facilities are established, the system maintains lower inventory levels but faces higher backlogs. The supply chain essentially prioritizes unmet demand over the high holding and spoilage costs associated with stored inventory.

Conclusion

This research provides a comprehensive framework for managing the complexities of perishable product supply chains. By linearizing FIFO constraints and utilizing decomposition methods, the authors transformed a non-convex problem into an efficient MILP. While the current social fairness index is a proof-of-concept based on poverty and unemployment, future work will explore more refined definitions of equity and advanced decomposition techniques to handle global-scale supply chains.

Summary updated

Abstract

Perishable consumer products like food, cosmetics, and household chemicals face challenges in supply chain management due to limited shelf life and uncertainties in demand and transportation. To address some of these issues, this work proposes a robust optimization framework for jointly optimizing facility allocation and inventory management. The framework determines optimal locations for distribution centers and their assigned customers, as well as inventory policies that minimize the total costs related to transportation, distribution, and storage under uncertain demand in a robust setting. Specifically, we develop a two-stage mixed-integer linear programming (MILP) model is that incorporates First-In-First-Out (FIFO) inventory policy to reduce spoilage. The bilinear FIFO constraints are linearized to improve computational efficiency. Social equity is integrated by defining a fairness index and incorporating it in facility allocation. Demand uncertainty is tackled using a robust optimization approach with affine demand functions to handle multiple scenarios. The model is solved using row and column generation techniques for scalability. Overall, this robust optimization framework is expected to enhance supply chain resilience, reduce waste, and improve cost-effectiveness in managing perishable products.
BibTeX
@inproceedings{naraghi2025joint,
  author = {Saba Ghasemi Naraghi and Zheyu Jiang},
  title = {{Joint Optimization of Fair Facility Allocation and Robust Inventory Management for Perishable Consumer Products}},
  booktitle = {Systems},
  year = {2025},
  doi = {10.69997/sct.153925},
}
Figure