Sustainable optimisation of Cold Food Supply Chains: a bi-objective inventory–routing model under demand uncertainty
Loading...
Files
Full Text E-thesis
Date
2025-12-01
Authors
Jahdi, Soodeh
Journal Title
Journal ISSN
Volume Title
Publisher
University College Cork
Published Version
Abstract
Cold Food Supply Chain (CFSC) plays a critical role in ensuring the availability, safety, and freshness of temperature-sensitive products, yet they remain particularly vulnerable to demand volatility, energy-intensive operations, and rising sustainability pressures. Maintaining precise temperature conditions throughout transportation and storage significantly increases energy consumption and operational cost, while also contributing to elevated Carbon Dioxide (CO2) emissions. These challenges are compounded by the fact that demand for perishable goods is frequently non-stationary and difficult to predict, with fluctuations driven by weather, promotions, consumer behaviour, and seasonal variations. Traditional replenishment policies such as the fixed-order (R, Q) approach lack the flexibility required in such uncertain environments and often produce inflated safety stocks and inefficient routing decisions. In cold chains, where excessive storage accelerates spoilage, increases energy use, and undermines product freshness, the consequences are particularly severe.
Responding to these challenges, this thesis develops an integrated, bi-objective Mixed-Integer Programming (MIP) model based on the adaptive (R, S) replenishment policy. The model jointly optimises replenishment timing, order-up-to levels, and vehicle routes in a multi-period distribution environment characterised by non-stationary stochastic demand. Two conflicting objectives are addressed simultaneously: (I) minimising total expected cost, and (II) minimising total expected CO2 emissions arising from transportation activity and refrigeration load. Given the combinatorial complexity of the multi-period Inventory Routing Problem (IRP) under uncertainty, the Augmented Epsilon Constraint (AEC) method is employed to generate Pareto-efficient solutions, supported by lexicographic optimisation to construct an accurate payoff table. All models are solved using IBM ILOG CPLEX, and solution robustness is assessed through 10,000-run Monte Carlo simulations.
The findings demonstrate that the (R, S) policy consistently outperforms the conventional (R, Q) policy across all examined demand patterns including constant, increasing, decreasing, life cycle, and erratic. In the case study, the (R, S) policy reduces total cost by 2.2% and CO2 emissions by 8.5%. Across 400 randomly generated problem instances, cost improvements grow with longer planning horizons and higher Coefficients of Variation (CV), highlighting the model’s effectiveness in volatile environments. The results also reveal that while modest emission reductions can be achieved with limited cost increases, deeper decarbonisation requires disproportionately higher investment, reflecting a clear non-linear trade-off along the Pareto frontier.
Overall, the thesis provides an analytic and practical framework for uncertainty-aware, energy-efficient cold-chain planning. The contributions support managers and logistics planners in making informed decisions that balance cost efficiency, operational responsiveness, and environmental responsibility, ultimately fostering more sustainable and resilient CFSC.
Description
Keywords
Cold Food Supply Chain Management , Inventory Routing Problem , Sustainable supply chain management , Food supply chain optimisation , Multi-objective Optimisation , Demand uncertainty , Data analytics , Perishable food logistics
Citation
Jahdi, S. 2025. Sustainable optimisation of Cold Food Supply Chains: a bi-objective inventory–routing model under demand uncertainty. MRes Thesis, University College Cork.