Haozhe Mao

Mr. Haozhe Mao


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
9 May 2026 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MJMei-xian Jiang
Mei-xian Jiang
FWFan Wu
Fan Wu
HMHaozhe Mao
Haozhe Mao
GBGuanjun BAO
Guanjun BAO
FWFan Wu
Fan Wu
Under the background of green and low-carbon port development, the single-terminal continuous berth dynamic scheduling problem is simultaneously affected by multiple factors, including dynamic vessel arrivals, tidal conditions, quay crane resources, and carbon emission costs, and has become a complex decision-making problem that must balance operational efficiency and low-carbon objectives. To address this issue, the problem is characterized in this study as an event-driven, dynamic spatiotemporal resource allocation and cost-coordinated optimization problem. A dynamic berth scheduling model is established with the objective of minimizing the total cost composed of waiting cost, delay cost, and carbon emission cost, and the problem is further formulated as a Markov decision process. Considering that conventional experience replay mechanisms are insufficient in exploiting critical samples and often suffer from limited training stability in complex dynamic scenarios, this study introduces an experience quality evaluation mechanism and a dual-replay-buffer collaborative training strategy on the basis of the DDQN algorithm, thereby proposing the QER-DDQN algorithm. The experimental results indicate that, under highly complex and highly congested test scenarios, QER-DDQN shows relatively better cost-control performance, with the average total cost reduced by approximately 8.08%, 10.77%, 7.78%, and 2.43% compared with FCFS, GA, PER-DDQN, and DDQN, respectively. The ablation study and scheduling scheme analysis further demonstrate that the experience quality evaluation mechanism and the dual-replay-buffer collaborative training strategy help improve the utilization efficiency of critical samples and, to a certain extent, enhance the local shoreline utilization structure, thereby alleviating the accumulation of waiting time and the propagation of delays caused by local resource conflicts. Further low-carbon scenario analysis shows that carbon tax level and shore power coverage rate both affect berth scheduling outcomes and carbon emission performance, and that appropriate low-carbon policy constraints and shore power infrastructure configuration are conducive to promoting the coordinated optimization of port scheduling efficiency and emission reduction objectives.

13 October 2026 in Discover Applied Sciences
Verified
Journal Article
Open access
PDF file
MJMei-xian Jiang
Mei-xian Jiang
YGYao Gu
Yao Gu
HMHaozhe Mao
Haozhe Mao
WTWenyao Tong
Wenyao Tong
HZHuawen Zhang
Huawen Zhang
GWGuanghua Wu
Guanghua Wu
Reliable online inspection of playing card surfaces is difficult because small, low-contrast defects are easily confused with dense printed graphics. This study proposes a defect-aware YOLOv8n framework integrating a ResNet50 backbone, DeCA, an improved BiFPN with SPD-Conv, a dynamic adaptive decoupled head, and Wise-IoU v3. These components strengthen weak-feature extraction, suppress background interference, preserve multiscale details, and improve localization. A production dataset of 2750 original images covering edge defect, ink missing, fold, and stain was augmented to 6000 images and divided 7:1:2 at the source-image level to prevent information leakage. Across three independent training runs, the proposed model achieved $$96.27\pm 0.16\%$$ mAP@50 and $$77.07\pm 0.18\%$$ mAP@50:95, corresponding to mean improvements of 12.77 and 11.64 percentage points over YOLOv8n, respectively. It also reached 97 FPS with 21.85 M parameters on the experimental GPU, indicating a favorable detector-level accuracy-speed balance under the evaluated hardware conditions. The evaluation is restricted to one production-line dataset; no external or cross-domain validation was conducted, and no generalization or transferability claim is made.