Answers are grounded in this paper and cite their sources.
Optimization study demonstrates multi-objective stochastic capacity configuration cuts unmet-demand risk by 42% in air-cargo hub terminals, indicating that cargo-mix volatility drives bottlenecks.
Capacity planning for air-cargo hub terminals is still performed stage by stage against deterministic mean-value forecasts, which conceals both the coupling of the six-stage handling chain and the pronounced upper-tail variability of cargo demand. This study proposes a risk-averse configuration framework that jointly decides line deployment, cargo-class assignment, and capacity tier across receiving, screening, sortation, storage, ULD build-up, and loading. The model is a two-stage stochastic program optimizing four criteria simultaneously—time-weighted service level, life-cycle cost, a Pollaczek–Khinchine congestion penalty, and a CVaR-augmented unmet-demand measure—with chance constraints handled by sample-average approximation and 5000 Monte Carlo scenarios compressed to 30 by forward Kantorovich reduction. An improved NSGA-II with hierarchical function–class–tier encoding, sequence-aware repair, time-priority mutation, and scenario-recursive fitness evaluation solves the resulting mixed-integer problem. On a calibrated super-hub instance (2800 t outbound per day), the knee-point design attains 96.4% of the service-maximal service level at 88.0% of its cost and cuts the conditional value-at-risk of unmet demand at the 0.90 confidence level by about 42% versus a deterministic-equivalent plan; the value of the stochastic solution reaches 9.9%. The algorithm outperforms NSGA-II, NSGA-III, MOEA/D-DE, SPEA2, and MOPSO on hypervolume and inverted generational distance. Compositional volatility, not aggregate volume, emerges as the dominant driver of service collapse.
No takes yet. Share an insight, caveat, or question.
吕凤禄 et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: