Liying Wang

Research Interests

• Operations Research, Game Theory
• Multi-agent systems
• Reinforcement Learning, AI-Based Methods
• AI-enabled robots, Autopilot
• Generative Adversarial Networks (GAN)
• Counterfactual Regret Minimization (CFR)
• Countinual Learning

Publications

Paper

[1] Zekun Duan, Genjiu Xu, Xin Liu, Jiayuan Ma, **Liying Wang**, “Optimal confrontation position selecting games model and its application to one-on-one air combat”, Defence Technology, 2023.

[2] Mengda Ji, Genjiu Xu, Zekun Duan, Liying Wang, Zesheng Li, Jianjun Ge, Mingqiang Li, “Cooperative Pursuit with Multiple Pursuers based on Deep Minimax Q-learning”, Aerospace Science and Technology, 2023. (TOP)

[3] Liying Wang,Yi Liu, Zesheng Li, Xin Liu, Genjiu Xu, “Multi-UAV Cooperative Target Allocation Method Based on Auction Mechanism”, Journal of Command and Control, 2023. [Accepted]

[4] Liying Wang, Genjiu Xu, Mengda Ji,”A CFR-based method for Solving Mobile Robot Confrontation Problems”, IEEE Robotics and Automation Letters, 2023.[Under review]

Conference

[1] Zesheng Li, Genjiu Xu, Liying Wang, “A Generalizable Autumous Maneuvering Decesion Making Method for UCAV Air Combat Combing PER-D3QN and Zero-Sum Markov Game” in International Conference on Autonomous Unmanned Systems，2023.

Research Experience

Reinforcement Learning and Subgame Solving for Incomplete Information Pursuit-Evasion Game (2022/11 – 2023/06)

• Discretized the pursuit-evasion game dynamics and action space, representing simultaneous action selection using information sets.
• Trained a neural network using reinforcement learning to evaluate state values.
• Utilized the CFR-D algorithm to solve for equilibrium in extended-form Subgames.

Cooperative Task Assignment for Multi-Agent System (2022/03 – 2022/09)

• Developed an optimal model for cooperative task assignment under time window constraints.
• Proposed the CBAA-TWC algorithm, a consensus-based auction algorithm addressing task assignment under time window constraints.
• Implemented a discrete time-varying price-increasing auction mechanism with individual rationality.

Reinforcement Learning for Dynamic Fire Tracking (2021/09 – 2022/03)

• Modeled fire propagation, quadcopter movement, and observations.
• Designed a reward function considering observation coverage and fire development trends.
• Employed mean-field representation to capture communication interactions among multiple agents and applied reinforcement learning to learn the Q-function.

Education

Northwestern Polytechnical University (Xi’an, CHINA)

• Master of Science in Electronic Information (2021/09 – 2024/03(expected))
• Advisor: Prof. Genjiu Xv
• The First Prize Scholarship
• GPA: 86/100
• Courses: Mathematical Statistics, Game Theory and Mechanism Design, Machine Learning and Model Optimization, Intelligent Information Processing and Collaboration for Unmanned Systems, etc.

NanJing Tech University (Nanjing, CHINA)

• Bachelor of Science in Information and Computational Science (2017/09 – 2021/06)
• National Encouragement Scholarships and First-Class Scholarships for THREE consecutive years
• Pacemaker to Merit Student; Nanjing Marathon Medical Volunteers
• GPA: 3.85/4.0
• Ranking: 1/59
• Courses: Mathematical Analysis, Advanced Algebra, Probability Theory, Operations and Optimization, Numerical Analysis, Analytic Geometry, Data Structure and Algorithm, Pattern Recognition, etc.

Skill

• Programming Languages: Python, Java, C++
• Software: MATLAB, TensorFlow, PyTorch
• Programming Practice: Reinforcement Learning (DQN, PPO, MADDPG), Deep Learning (AlexNet, Transformer, GAN), CFR Algorithm (CFR, MCCFR, CFR-D)

Award

• The Second Prize, Shaanxi Province 2022 Intelligent Unmanned System Application Challenge (2022/09)
• The Second Prize, Jiangsu Province Data Mining Challenge (2020/09)

Else

• Blog: https://zuti666.github.io
• GitHub: https://github.com/zuti666
• Language: English (IELTS: 6.5)

Contact

• Email: li.ying.wang@outlook.com
• phone: (+86) 187-616-06609