A stochastic multi-objective optimization framework for real-time IoT resource allocation in urban digital twins
2026, vol.18 , no.4, pp. 3-10
Article [2026-04-01]
Urban Digital Twins (UDTs) synchronize virtual city replicas with real-time IoT data, yet unrestricted ingestion exhausts bandwidth, computation and energy budgets while inflating latency. The paper formulates UDT data ingestion as a stochastic multi-objective integer program with chance-constrained bandwidth, energy and latency risk, using a redundancy-aware submodular Mutual-Information utility objective, and solves it with the Optimized Adaptive Data Ingestion (OADI) framework - a simheuristic coupling Monte-Carlo scenario generation with the novel Stochastic Adaptive Reference Point Mechanism (SARPM), which reshapes NSGA-III reference points from the posterior constraint-violation distribution. On PeMS and OpenAQ, OADI achieves a 34 higher Information-to-Cost Ratio than the best greedy heuristic, a 94.8 constraint-satisfaction rate (estimated on 1,000 held-out scenarios), and statistically significant Hypervolume and IGD+ improvements over seven baselines and a SARPM ablation (Wilcoxon p < 0.01).
urban digital twins, IoT resource allocation, stochastic multi-objective optimization, NSGA-III, adaptive data ingestion
https://doi.org/10.59035/QKLU3734
Mohamed M. H. Maatouk, Mhd Walid Koubeisi, Mahmoud Mohamed. A stochastic multi-objective optimization framework for real-time IoT resource allocation in urban digital twins. International Journal on Information Technologies and Security, vol.18 , no.4, 2026, pp. 3-10. https://doi.org/10.59035/QKLU3734