Many objective test suite for perishable goods distribution with product-specific perishability models

Abstract

Perishable goods distribution (PGD) systems operate under tightly coupled routing, scheduling, refrigeration, and product-quality constraints, making them among the most challenging classes of multi-objective vehicle routing problems. Despite the growing application of Multi-Objective Evolutionary Algorithms (MOEAs) to cold-chain logistics, the absence of benchmark test suites that simultaneously capture the structural complexity of real-world perishable goods networks continues to limit rigorous algorithm design, evaluation and fair comparative analysis. Existing routing benchmarks often neglect critical characteristics such as heterogeneous perishability behavior, asymmetric spatial clustering, time-critical delivery framework, route-position-dependent quality degradation, and product incompatibility constraints, resulting in simplified optimization landscapes that insufficiently model practical cold-chain operations. This study proposes a structured multi-objective benchmark test suite with 15 PGD instances generated through a customer-type-driven framework with explicitly controlled structural properties and difficulty characteristics to address this gap. The proposed test suite integrates demand heterogeneity, temporal feasibility pressure, multi-source incompatibility relationships, specific quality loss dynamics, and diverse spatial topologies to evaluate the performance of MOEAs under realistic cold-chain conditions. The representative state-of-the-art MOEAs are evaluated across conflicting objectives and constraints. Experimental analyses and results show performance variations across instances, thereby providing a scalable and structurally diverse benchmark test suite with baseline performance. This study provides a foundation for the development, evaluation, and reproducible comparison of future optimization methods for sustainable and efficient PGD systems.