Huan Xia
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Huan Xia has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Hao Tang
Civil Aviation Electronic Technology Co., Ltd., The Second Research Institute of Civil Aviation Administration of China, Chengdu 610041, China
1 shared publication - Jun Xu
Civil Aviation Electronic Technology Co., Ltd., The Second Research Institute of Civil Aviation Administration of China, Chengdu 610041, China
1 shared publication - Lishun Zhang
School of Computer and Software Engineering, Xihua University, Chengdu 610039, China
1 shared publication - Ping Zhang
Civil Aviation Electronic Technology Co., Ltd., Chengdu 611400, China
1 shared publication - Qian Luo
Second Research Institute of the Civil Aviation Administration of China, Chengdu 610041, China
1 shared publication - Tao Wen
Civil Aviation Electronic Technology Co., Ltd., Chengdu 611400, China
1 shared publication
Latest Publications
- Open Access
- 6 August 2026
- Published by MDPI AG in Algorithms
- Vol. 19 (8) , 651
An Integrated Scheduling Model for Airport Apron-Bus Drivers Under Stochastic Demand
Airport apron buses, which shuttle passengers between terminal gates and remotely parked aircraft, are vital for efficient ground operations. However, flight delays and disruptions make demand for apron-bus services uncertain and time-varying, creating substantial challenges for planning driver capacity and work schedules. To tackle this challenge, we develop an Integrated Stochastic-Flexible Planning Model (ISFPM), formulated as a mixed-integer linear program (MILP), that simultaneously optimizes workforce sizing, duty scheduling, and roster assignment for apron-bus drivers. The objective is to minimize the sum of labor costs and the expected penalty for understaffing. This penalty is evaluated under the assumption that driver demand in each period follows a Poisson-binomial distribution, which is derived from probabilistic models of flight delays. For computational efficiency, we reformulate the expected penalty term using continuity-corrected normal approximations based on the cumulative distribution function (CDF). Furthermore, we incorporate practical workforce flexibility features, including hourly-granularity duty start times and heterogeneous workday patterns across roster groups. The computational study is based on Beijing Capital International Airport and accompanied by deidentified replication materials. Across 100 materialized baseline scenarios, the ISFPM uses 148 drivers instead of the 175-driver deterministic fixed-shift benchmark, reducing average management cost by 21.9% and passenger waiting time by 85.6%. Across 18 representative-day scenarios with correlated and severe disruptions, it reduces mean management cost by 15.8% and passenger waiting time by 53.6%. The disruption-scenario analysis also indicates that deterministic flexible staffing attains the lowest waiting time at a higher cost, while common apron travel-time shocks reduce the service advantage of the ISFPM.
- 6 August 2026
- Published by MDPI AG in Algorithms
- Vol. 19 (8) , 651
Open AccessAn Integrated Scheduling Model for Airport Apron-Bus Drivers Under Stochastic Demand
Airport apron buses, which shuttle passengers between terminal gates and remotely parked aircraft, are vital for efficient ground operations. However, flight delays and disruptions make demand for apron-bus services uncertain and time-varying, creating substantial challenges for planning driver capacity and work schedules. To tackle this challenge, we develop an Integrated Stochastic-Flexible Planning Model (ISFPM), formulated as a mixed-integer linear program (MILP), that simultaneously optimizes workforce sizing, duty scheduling, and roster assignment for apron-bus drivers. The objective is to minimize the sum of labor costs and the expected penalty for understaffing. This penalty is evaluated under the assumption that driver demand in each period follows a Poisson-binomial distribution, which is derived from probabilistic models of flight delays. For computational efficiency, we reformulate the expected penalty term using continuity-corrected normal approximations based on the cumulative distribution function (CDF). Furthermore, we incorporate practical workforce flexibility features, including hourly-granularity duty start times and heterogeneous workday patterns across roster groups. The computational study is based on Beijing Capital International Airport and accompanied by deidentified replication materials. Across 100 materialized baseline scenarios, the ISFPM uses 148 drivers instead of the 175-driver deterministic fixed-shift benchmark, reducing average management cost by 21.9% and passenger waiting time by 85.6%. Across 18 representative-day scenarios with correlated and severe disruptions, it reduces mean management cost by 15.8% and passenger waiting time by 53.6%. The disruption-scenario analysis also indicates that deterministic flexible staffing attains the lowest waiting time at a higher cost, while common apron travel-time shocks reduce the service advantage of the ISFPM.
- Open Access
- 1 July 2026
- Published by MDPI AG in Electronics
- Vol. 15 (13) , 2867
Multi-Statistic Disentangled LSTM with Hidden-State Feature Extraction for Aero-Engine Remaining Useful Life Prediction
Accurate remaining useful life (RUL) prediction for aero-engines is important for condition-based maintenance and safety-oriented health management. Long short-term memory (LSTM) networks are widely used for this task, but two limitations remain important in multi-sensor degradation modeling: hidden states generated over the full window are often under-utilized, and attention mechanisms may overemphasize locally fluctuating sensor readings. This paper proposes a Multi-Statistic Disentangled LSTM (MSD-LSTM) framework for aero-engine RUL prediction. The framework first applies Savitzky–Golay filtering to smooth high-frequency signal fluctuations. A hidden-state feature extraction module then combines feature-level disentangled extraction and Global Average Pooling to use the LSTM hidden-state sequence beyond the final recurrent output. In parallel, a Multi-Statistic Pooler summarizes each input window using minimum, maximum, standard deviation, and mean statistics, and its output is fused with a self-attention branch through a static-gating mechanism. On the NASA C-MAPSS benchmark, MSD-LSTM achieves RMSE values of 10.45 and 12.33 on FD001 and FD002, respectively, and ranks first in RMSE on three of the four sub-datasets and first in SCORE on two sub-datasets among the compared recent methods. Ablation and fusion analyses show that both the hidden-state extraction and statistic-guided fusion components contribute to stable RUL prediction.
- 1 July 2026
- Published by MDPI AG in Electronics
- Vol. 15 (13) , 2867
Open AccessMulti-Statistic Disentangled LSTM with Hidden-State Feature Extraction for Aero-Engine Remaining Useful Life Prediction
Accurate remaining useful life (RUL) prediction for aero-engines is important for condition-based maintenance and safety-oriented health management. Long short-term memory (LSTM) networks are widely used for this task, but two limitations remain important in multi-sensor degradation modeling: hidden states generated over the full window are often under-utilized, and attention mechanisms may overemphasize locally fluctuating sensor readings. This paper proposes a Multi-Statistic Disentangled LSTM (MSD-LSTM) framework for aero-engine RUL prediction. The framework first applies Savitzky–Golay filtering to smooth high-frequency signal fluctuations. A hidden-state feature extraction module then combines feature-level disentangled extraction and Global Average Pooling to use the LSTM hidden-state sequence beyond the final recurrent output. In parallel, a Multi-Statistic Pooler summarizes each input window using minimum, maximum, standard deviation, and mean statistics, and its output is fused with a self-attention branch through a static-gating mechanism. On the NASA C-MAPSS benchmark, MSD-LSTM achieves RMSE values of 10.45 and 12.33 on FD001 and FD002, respectively, and ranks first in RMSE on three of the four sub-datasets and first in SCORE on two sub-datasets among the compared recent methods. Ablation and fusion analyses show that both the hidden-state extraction and statistic-guided fusion components contribute to stable RUL prediction.