Static resource-allocation models can overlook spatial inequity and changing urban demand. We propose a closed-loop, online spatiotemporal multi-objective framework that jointly addresses equity and efficiency in smart city infrastructure. A CNN-BiGRU forecaster with additive attention estimates demand means and variances; a self-adaptive estimation-of-distribution evolutionary optimizer searches for allocation policies along the Pareto frontier of spatial accessibility inequality and service coverage minus cost. A sliding-window Kolmogorov-Smirnov detector identifies distribution shifts and triggers re-optimization only in affected subregions. Finally, a decision-support layer selects a policy from the Pareto set using an equity-adjusted social welfare function combining an Atkinson-based term and a Gini penalty. Predictions, allocation decisions, and observed service outcomes form a feedback loop. Evaluation on synthetic demand shifts and New York City taxi records, as reported in this study, compares forecasting, solution quality, equity, and computation time with established baselines. The framework offers an integrated approach to uncertainty-aware forecasting, adaptive resource allocation, and transparent equity-efficiency trade-offs.

Xinyi Pan. Dynamic Equity-Aware Pareto Optimization for Predictive Resource Allocation in Smart City Infrastructure. Academic Journal of Computing & Information Science (2026), Vol. 9, Issue 8: 58-68. https://doi.org/10.25236/AJCIS.2026.090808.

[1] Chen, L. (2025). Smart cities and social equity: A review of digital urban governance in southeast asia. International Journal of Social Science and Applied Technology.

[2] Haque, M., Beckers, D., Costales, E., Aad, S., Sharifi, A., et al. (2025). A systematic review of research on just, equitable, responsible, and inclusive smart cities. Technology in Society.

[3] Zhang, F. (2024). Evolutionary algorithm for dynamic resource allocation and its applications. International Journal of Simulation Modelling.

[4] Yang, B., Sun, S., Li, J., Lin, X., & Tian, Y. (2019). Traffic flow prediction using LSTM with feature enhancement. Neurocomputing.

[5] Shi, X., Chen, Z., Wang, H., Yeung, D., et al. (2015). Convolutional LSTM network: A machine learning approach for precipitation nowcasting. Advances in Neural Information Processing Systems.

[6] Zhang, J., Zheng, Y., & Qi, D. (2017). Deep spatio-temporal residual networks for citywide crowd flows prediction. Proceedings of the AAAI Conference on Artificial Intelligence.

[7] Xie, H., Ding, R., Qiao, G., Dai, C., & Bai, W. (2024). Research on ship traffic flow prediction using CNN-BiGRU and WOA with multi-objective optimization. IEEE Access.

[8] Deb, K., Pratap, A., Agarwal, S., et al. (2002). A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation.

[9] Zhang, Q., & Li, H. (2007). MOEA/d: A multiobjective evolutionary algorithm based on decomposition. IEEE Transactions on Evolutionary Computation.

[10] Songserm, N., Pitakaso, R., Srichok, T., Khonjun, S., et al. (2026). Equity-preserving public health resource allocation using multi-objective safe reinforcement learning: Evidence from thailand. International Journal of Environmental Research and Public Health.

[11] Dębski, R., & Dreżewski, R. (2024). Multi-objective ship route optimisation using estimation of distribution algorithm. Applied Sciences.

[12] Zhou, S., Wang, D., Liu, Z., Ma, Y., & Liu, X. (2025). Equity-oriented multi-objective optimization allocation strategies for urban water resources. Water.

[13] Weber, T. (2026). Fair allocation of an exhaustible resource. 2026 European Control Conference (ECC).

[14] TLC, N. (2022). NYC taxi and limousine commission trip record data. 2022.

[15] Bureau, U. C. (2008). Income, earnings, and poverty data from the... American community survey. books.google.com.

[16] Blank, J., & Deb, K. (2020). Pymoo: Multi-objective optimization in python. Ieee Access.