Forecast-Driven Dispatch for Distributed Service Fleets

Objective. We present an applied machine-learning framework for scheduling cleaning crews across distributed public-restroom fleets, replacing industry-standard fixed-interval and threshold-triggered rules with a predict-then-optimize pipeline that couple’s usage forecasting, a latent dirtiness state model, and travel-aware preemptive dispatch.

Methods. On a calibrated synthetic city of 50 restroom units across five location types — using a non-homogeneous Poisson usage generator with weather and event bursts and a detour-scaled Euclidean travel matrix — we benchmark five machine-learning contenders (seasonal-naive + greedy, LightGBM quantile forecaster + greedy, Cox proportional-hazards + greedy, Cox PH + OR-Tools mini-VRP, and a rolling-horizon variant) against fixed-interval and per-type usage-threshold baselines, plus a perfect-future-usage oracle. All numbers are produced by a single reproducible script (6 seeds, 2-week train / 4-week evaluation, with 95% confidence intervals and paired -tests).

Results. At the nominal crew budget (220 h/wk, 3 crews), LGBM+Greedy reduced hours above threshold per unit per week from 19.4 (FI-12h) to 1.83 — a 91% reduction (, paired) — using only 79% of the budget. A 3-hour oracle matches LGBM+Greedy within noise, indicating that short-horizon usage prediction is saturated at this fleet scale. Cox proportional-hazards variants under-perform the fixed-interval baseline. A generator–model mismatch stress test (Hawkes-style tails,  drift, spill shocks) preserves the ordering.

Conclusion. A simple usage-forecast-driven greedy dispatcher dominates more sophisticated routing-based alternatives at this fleet scale and provides an engineering baseline for a real-world pilot. A complete reproducible Python implementation accompanies this manuscript as supplementary material.

predictive dispatch, service fleet scheduling, LightGBM, survival analysis, vehicle routing, predict-then-optimize

Agrawal, A., Amos, B., Barratt, S., Boyd, S., Diamond, S., & Kolter, J. Z. (2019). Differentiable convex optimization layers. Advances in Neural Information Processing Systems, 32.

Alfonso-Lizarazo, E. H., Montoya-Torres, J. R., & Gutiérrez-Franco, E. (2013). Modeling reverse logistics process in the agro-industrial sector: The case of the palm oil supply chain. Applied Mathematical Modelling, 37(23), 9652–9664. https://doi.org/10.1016/j.apm.2013.05.015

Amos, B., & Kolter, J. Z. (2017). OptNet: Differentiable optimization as a layer in neural networks. Proceedings of the 34th International Conference on Machine Learning, 136–145.

Anh Khoa, T., Phuc, C. H., Lam, P. D., Nhu, L. M. B., Trong, N. M., & Phuong, N. T. H. (2020). Waste management system using IoT-based machine learning in university. Wireless Communications and Mobile Computing, 2020, Article 6138637. https://doi.org/10.1155/2020/6138637

Beliën, J., & Demeulemeester, E. (2008). A branch-and-price approach for integrating nurse and surgery scheduling. European Journal of Operational Research, 189(3), 652–668. https://doi.org/10.1016/j.ejor.2006.10.060

Carvalho, T. P., Soares, F. A. A. M. N., Vita, R., Francisco, R. da P., Basto, J. P., & Alcalá, S. G. S. (2019). A systematic literature review of machine learning methods applied to predictive maintenance. Computers & Industrial Engineering, 137, Article 106024. https://doi.org/10.1016/j.cie.2019.106024

Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794. https://doi.org/10.1145/2939672.2939785

Cox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society, Series B, 34(2), 187–220. https://doi.org/10.1111/j.2517-6161.1972.tb00899.x

Dasaklis, T. K., & Pappis, C. P. (2013). Supply chain management in view of climate change: An overview of possible impacts and the road ahead. Journal of Industrial Engineering and Management, 6(4), 1139–1161. https://doi.org/10.3926/jiem.883

Davidson-Pilon, C. (2019). lifelines: Survival analysis in Python. Journal of Open Source Software, 4(40), Article 1317. https://doi.org/10.21105/joss.01317

Donti, P. L., Amos, B., & Kolter, J. Z. (2017). Task-based end-to-end model learning in stochastic optimization. Advances in Neural Information Processing Systems, 30, 5484–5494.

Dragomir, A. G., & Doerner, K. F. (2024). Waste collection routing: A survey on problems and methods. Central European Journal of Operations Research, 32(2), 399–434. https://doi.org/10.1007/s10100-023-00892-y

Elmachtoub, A. N., & Grigas, P. (2022). Smart “predict, then optimize”. Management Science, 68(1), 9–26. https://doi.org/10.1287/mnsc.2020.3922

Faccio, M., Persona, A., & Zanin, G. (2011). Waste collection multi-objective model with real-time traceability data. Waste Management, 31(12), 2391–2405. https://doi.org/10.1016/j.wasman.2011.07.005

Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 29(5), 1189–1232. https://doi.org/10.1214/aos/1013203451

Gneiting, T., & Raftery, A. E. (2007). Strictly proper scoring rules, prediction, and estimation. Journal of the American Statistical Association, 102(477), 359–378. https://doi.org/10.1198/016214506000001437

Hawkes, A. G. (1971). Spectra of some self-exciting and mutually exciting point processes. Biometrika, 58(1), 83–90. https://doi.org/10.1093/biomet/58.1.83

Jardine, A. K. S., Lin, D., & Banjevic, D. (2006). A review on machinery diagnostics and prognostics implementing condition-based maintenance. Mechanical Systems and Signal Processing, 20(7), 1483–1510. https://doi.org/10.1016/j.ymssp.2005.09.012

Johansson, O. M. (2006). The effect of dynamic scheduling and routing in a solid waste management system. Waste Management, 26(8), 875–885. https://doi.org/10.1016/j.wasman.2005.09.004

Kalbfleisch, J. D., & Prentice, R. L. (2002). The statistical analysis of failure time data (2nd ed.). Wiley. https://doi.org/10.1002/9781118032985

Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q., & Liu, T.-Y. (2017). LightGBM: A highly efficient gradient boosting decision tree. Advances in Neural Information Processing Systems, 30, 3146–3154.

Koenker, R., & Bassett, G. (1978). Regression quantiles. Econometrica, 46(1), 33–50. https://doi.org/10.2307/1913643

Lim, B., Arık, S. Ö., Loeff, N., & Pfister, T. (2021). Temporal fusion transformers for interpretable multi-horizon time series forecasting. International Journal of Forecasting, 37(4), 1748–1764. https://doi.org/10.1016/j.ijforecast.2021.03.012

Makridakis, S., Spiliotis, E., & Assimakopoulos, V. (2022a). M5 accuracy competition: Results, findings, and conclusions. International Journal of Forecasting, 38(4), 1346–1364. https://doi.org/10.1016/j.ijforecast.2021.11.013

Makridakis, S., Spiliotis, E., & Assimakopoulos, V. (2022b). The M5 competition: Background, organization, and implementation. International Journal of Forecasting, 38(4), 1325–1336. https://doi.org/10.1016/j.ijforecast.2021.07.007

Mandi, J., & Guns, T. (2020). Interior point solving for LP-based prediction+optimisation. Advances in Neural Information Processing Systems, 33, 7272–7282.

Markov, I., Bierlaire, M., Cordeau, J.-F., Maknoon, Y., & Varone, S. (2020). Waste collection inventory routing with non-stationary stochastic demands. Computers & Operations Research, 113, Article 104798. https://doi.org/10.1016/j.cor.2019.104798

Mes, M., Schutten, M., & Pérez Rivera, A. (2014). Inventory routing for dynamic waste collection. Waste Management, 34(9), 1564–1576. https://doi.org/10.1016/j.wasman.2014.05.011

Morais, C. S. de, Ramos, T. R. P., Lopes, M., & Barbosa-Póvoa, A. P. (2024). A data-driven optimization approach to plan smart waste collection operations. International Transactions in Operational Research, 31(4), 2178–2208. https://doi.org/10.1111/itor.13235

Nuortio, T., Kytöjoki, J., Niska, H., & Bräysy, O. (2006). Improved route planning and scheduling of waste collection and transport. Expert Systems with Applications, 30(2), 223–232. https://doi.org/10.1016/j.eswa.2005.07.009

Ogata, Y. (1988). Statistical models for earthquake occurrences and residual analysis for point processes. Journal of the American Statistical Association, 83(401), 9–27. https://doi.org/10.1080/01621459.1988.10478560

Ogata, Y. (1998). Space-time point-process models for earthquake occurrences. Annals of the Institute of Statistical Mathematics, 50(2), 379–402. https://doi.org/10.1023/A:1003403601725

Omi, T., Ueda, N., & Aihara, K. (2019). Fully neural network based model for general temporal point processes. Advances in Neural Information Processing Systems, 32, 2120–2129.

Oreshkin, B. N., Carpov, D., Chapados, N., & Bengio, Y. (2020). N-BEATS: Neural basis expansion analysis for interpretable time series forecasting. In Proceedings of the 8th International Conference on Learning Representations (ICLR 2020).

Perron, L., & Furnon, V. (2025). OR-Tools (Version 9.x) [Computer software]. Google.

Pillac, V., Gendreau, M., Guéret, C., & Medaglia, A. L. (2013). A review of dynamic vehicle routing problems. European Journal of Operational Research, 225(1), 1–11. https://doi.org/10.1016/j.ejor.2012.08.015

Psaraftis, H. N., Wen, M., & Kontovas, C. A. (2016). Dynamic vehicle routing problems: Three decades and counting. Networks, 67(1), 3–31. https://doi.org/10.1002/net.21628

Ramos, T. R. P., de Morais, C. S., & Barbosa-Póvoa, A. P. (2018). The smart waste collection routing problem: Alternative operational management approaches. Expert Systems with Applications, 103, 146–158. https://doi.org/10.1016/j.eswa.2018.03.001

Salinas, D., Flunkert, V., Gasthaus, J., & Januschowski, T. (2020). DeepAR: Probabilistic forecasting with autoregressive recurrent networks. International Journal of Forecasting, 36(3), 1181–1191. https://doi.org/10.1016/j.ijforecast.2019.07.001

Serradilla, O., Zugasti, E., Rodriguez, J., & Zurutuza, U. (2022). Deep learning models for predictive maintenance: A survey, comparison, challenges and prospects. Applied Intelligence, 52, 10934–10964. https://doi.org/10.1007/s10489-021-03004-y

Shchur, O., Türkmen, A. C., Januschowski, T., & Günnemann, S. (2021). Neural temporal point processes: A review. Proceedings of the 30th International Joint Conference on Artificial Intelligence (IJCAI 2021), 4585–4593. https://doi.org/10.24963/ijcai.2021/623

Si, X.-S., Wang, W., Hu, C.-H., & Zhou, D.-H. (2011). Remaining useful life estimation: A review on the statistical data-driven approaches. European Journal of Operational Research, 213(1), 1–14. https://doi.org/10.1016/j.ejor.2010.11.018

Solomon, M. M. (1987). Algorithms for the vehicle routing and scheduling problems with time window constraints. Operations Research, 35(2), 254–265. https://doi.org/10.1287/opre.35.2.254

Toth, P., & Vigo, D. (Eds.). (2014). Vehicle routing: Problems, methods, and applications (2nd ed., Vol. 18). MOS-SIAM Series on Optimization. https://doi.org/10.1137/1.9781611973594

Vicentini, F., Giusti, A., Rovetta, A., Fan, X., He, Q., Zhu, M., & Liu, B. (2009). Sensorized waste collection container for content estimation and collection optimization. Waste Management, 29(5), 1467–1472. https://doi.org/10.1016/j.wasman.2008.10.017

Zhang, W., Yang, D., & Wang, H. (2019). Data-driven methods for predictive maintenance of industrial equipment: A survey. IEEE Systems Journal, 13(3), 2213–2227. https://doi.org/10.1109/JSYST.2019.2905565

Zonta, T., da Costa, C. A., da Rosa Righi, R., de Lima, M. J., da Trindade, E. S., & Li, G. P. (2020). Predictive maintenance in the Industry 4.0: A systematic literature review. Computers & Industrial Engineering, 150, Article 106889. https://doi.org/10.1016/j.cie.2020.106889

Andrade, L. A. C. G. de, & Cunha, C. B. (2023). Disaggregated retail forecasting: A gradient boosting approach. Applied Soft Computing, 141, Article 110283. https://doi.org/10.1016/j.asoc.2023.110283

Chen, D., Imdahl, C., Lai, D., & Van Woensel, T. (2026). A reinforcement learning approach for the dynamic vehicle routing and scheduling problem with stochastic request times and time-dependent, stochastic travel times. Transportation Research Part C: Emerging Technologies, 182, Article 105387. https://doi.org/10.1016/j.trc.2025.105387

Ribeiro, R., & Fanzeres, B. (2026). Integrated estimate-and-optimize decision trees learning for two-stage linear decision-making problems. European Journal of Operational Research, 329(2), 607–628. https://doi.org/10.1016/j.ejor.2025.08.048

Soeffker, N., Ulmer, M. W., & Mattfeld, D. C. (2024). Balancing resources for dynamic vehicle routing with stochastic customer requests. OR Spectrum, 46(2), 331–373. https://doi.org/10.1007/s00291-024-00747-1

Yang, J., Wang, X., & Luo, Z. (2024). Few-shot remaining useful life prediction based on meta-learning with deep sparse kernel network. Information Sciences, 653, Article 119795. https://doi.org/10.1016/j.ins.2023.119795

Yu, Q., Cao, Z., Wang, R., Yang, Z., Deng, L., Hu, M., Luo, Y., & Zhou, X. (2025). Dual-splitting conformal prediction for multi-step time series forecasting. Applied Soft Computing, 184, Article 113825. https://doi.org/10.1016/j.asoc.2025.113825