arXiv subjects

Shibo Chen

Publications and source records attributed to Shibo Chen.

13 recordsLinked to original sources

arXiv2026

Bottomonium suppression with a machine-learning-informed Debye mass

Motivated by recent progress in data-driven approaches, we introduce a machine-learning (ML)-informed Debye mass, extracted from lattice-informed inputs, exclusively in the complex-valued heavy-quark Kent State University (KSU) potential. The resulting complex potential is used to solve the real-time Schr\"odinger equation within the quantum trajectories (QTraj) framework for the evolution of bottomonium in the quark-gluon plasma. We then compute the nuclear modification factors and double ratios for bottomonium $\Upsilon(1S)$, $\Upsilon(2S)$, and $\Upsilon(3S)$ states in Pb-Pb collisions at $\sqrt{s_{NN}} = 5.02$ TeV. We compare our ML-induced results with those from the original KSU model and with experimental measurements from ALICE, ATLAS, and CMS collaborations. We find that the machine-learned Debye mass leads to improved agreement with data, particularly for excited states, highlighting the utility of machine learning in modeling in-medium QCD effects.

hep-ph↗

arXiv2026

M100: An Orchestrated Dataflow Architecture Powering General AI Computing

As deep learning-based AI technologies gain momentum, the demand for general-purpose AI computing architectures continues to grow. While GPGPU-based architectures offer versatility for diverse AI workloads, they often fall short in efficiency and cost-effectiveness. Various Domain-Specific Architectures (DSAs) excel at particular AI tasks but struggle to extend across broader applications or adapt to the rapidly evolving AI landscape. M100 is Li Auto's response: a performant, cost-effective architecture for AI inference in Autonomous Driving (AD), Large Language Models (LLMs), and intelligent human interactions, domains crucial to today's most competitive automobile platforms. M100 employs a dataflow parallel architecture, where compiler-architecture co-design orchestrates not only computation but, more critically, data movement across time and space. Leveraging dataflow computing efficiency, our hardware-software co-design improves system performance while reducing hardware complexity and cost. M100 largely eliminates caching: tensor computations are driven by compiler- and runtime-managed data streams flowing between computing elements and on/off-chip memories, yielding greater efficiency and scalability than cache-based systems. Another key principle was selecting the right operational granularity for scheduling, issuing, and execution across compiler, firmware, and hardware. Recognizing commonalities in AI workloads, we chose the tensor as the fundamental data element. M100 demonstrates general AI computing capability across diverse inference applications, including UniAD (for AD) and LLaMA (for LLMs). Benchmarks show M100 outperforms GPGPU architectures in AD applications with higher utilization, representing a promising direction for future general AI computing.

cs.LG↗

arXiv2026

GAS: Enhancing Reward-Cost Balance of Generative Model-assisted Offline Safe RL

Offline Safe Reinforcement Learning (OSRL) aims to learn a policy to achieve high performance in sequential decision-making while satisfying constraints, using only pre-collected datasets. Recent works, inspired by the strong capabilities of Generative Models (GMs), reformulate decision-making in OSRL as a conditional generative process, where GMs generate desirable actions conditioned on predefined reward and cost values. However, GM-assisted methods face two major challenges in OSRL: (1) lacking the ability to "stitch" optimal transitions from suboptimal trajectories within the dataset, and (2) struggling to balance reward targets with cost targets, particularly when they are conflict. To address these issues, we propose Goal-Assisted Stitching (GAS), a novel algorithm designed to enhance stitching capabilities while effectively balancing reward maximization and constraint satisfaction. To enhance the stitching ability, GAS first augments and relabels the dataset at the transition level, enabling the construction of high-quality trajectories from suboptimal ones. GAS also introduces novel goal functions, which estimate the optimal achievable reward and cost goals from the dataset. These goal functions, trained using expectile regression on the relabeled and augmented dataset, allow GAS to accommodate a broader range of reward-cost return pairs and achieve a better tradeoff between reward maximization and constraint satisfaction compared to human-specified values. The estimated goals then guide policy training, ensuring robust performance under constrained settings. Furthermore, to improve training stability and efficiency, we reshape the dataset to achieve a more uniform reward-cost return distribution. Empirical results validate the effectiveness of GAS, demonstrating superior performance in balancing reward maximization and constraint satisfaction compared to existing methods.

cs.LG↗

arXiv2026

Puzzle it Out: Local-to-Global World Model for Offline Multi-Agent Reinforcement Learning

Offline multi-agent reinforcement learning (MARL) aims to solve cooperative decision-making problems in multi-agent systems using pre-collected datasets. Existing offline MARL methods primarily constrain training within the dataset distribution, resulting in overly conservative policies that struggle to generalize beyond the support of the data. While model-based approaches offer a promising solution by expanding the original dataset with synthetic data generated from a learned world model, the high dimensionality, non-stationarity, and complexity of multi-agent systems make it challenging to accurately estimate the transitions and reward functions in offline MARL. Given the difficulty of directly modeling joint dynamics, we propose a local-to-global (LOGO) world model, a novel framework that leverages local predictions-which are easier to estimate-to infer global state dynamics, thus improving prediction accuracy while implicitly capturing agent-wise dependencies. Using the trained world model, we generate synthetic data to augment the original dataset, expanding the effective state-action space. To ensure reliable policy learning, we further introduce an uncertainty-aware sampling mechanism that adaptively weights synthetic data by prediction uncertainty, reducing approximation error propagation to policies. In contrast to conventional ensemble-based methods, our approach requires only an additional encoder for uncertainty estimation, significantly reducing computational overhead while maintaining accuracy. Extensive experiments across 8 scenarios against 8 baselines demonstrate that our method surpasses state-of-the-art baselines on standard offline MARL benchmarks, establishing a new model-based baseline for generalizable offline multi-agent learning.

cs.AI↗

arXiv2025

A Generative Model Enhanced Multi-Agent Reinforcement Learning Method for Electric Vehicle Charging Navigation

With the widespread adoption of electric vehicles (EVs), navigating for EV drivers to select a cost-effective charging station has become an important yet challenging issue due to dynamic traffic conditions, fluctuating electricity prices, and potential competition from other EVs. The state-of-the-art deep reinforcement learning (DRL) algorithms for solving this task still require global information about all EVs at the execution stage, which not only increases communication costs but also raises privacy issues among EV drivers. To overcome these drawbacks, we introduce a novel generative model-enhanced multi-agent DRL algorithm that utilizes only the EV's local information while achieving performance comparable to these state-of-the-art algorithms. Specifically, the policy network is implemented on the EV side, and a Conditional Variational Autoencoder-Long Short Term Memory (CVAE-LSTM)-based recommendation model is developed to provide recommendation information. Furthermore, a novel future charging competition encoder is designed to effectively compress global information, enhancing training performance. The multi-gradient descent algorithm (MGDA) is also utilized to adaptively balance the weight between the two parts of the training objective, resulting in a more stable training process. Simulations are conducted based on a practical area in Xi\'an, China. Experimental results show that our proposed algorithm, which relies on local information, outperforms existing local information-based methods and achieves less than 8\% performance loss compared to global information-based methods.

cs.LG↗

arXiv2024

Reinforcement Learning with Intrinsically Motivated Feedback Graph for Lost-sales Inventory Control

Reinforcement learning (RL) has proven to be well-performed and general-purpose in the inventory control (IC). However, further improvement of RL algorithms in the IC domain is impeded due to two limitations of online experience. First, online experience is expensive to acquire in real-world applications. With the low sample efficiency nature of RL algorithms, it would take extensive time to train the RL policy to convergence. Second, online experience may not reflect the true demand due to the lost sales phenomenon typical in IC, which makes the learning process more challenging. To address the above challenges, we propose a decision framework that combines reinforcement learning with feedback graph (RLFG) and intrinsically motivated exploration (IME) to boost sample efficiency. In particular, we first take advantage of the inherent properties of lost-sales IC problems and design the feedback graph (FG) specially for lost-sales IC problems to generate abundant side experiences aid RL updates. Then we conduct a rigorous theoretical analysis of how the designed FG reduces the sample complexity of RL methods. Based on the theoretical insights, we design an intrinsic reward to direct the RL agent to explore to the state-action space with more side experiences, further exploiting FG's power. Experimental results demonstrate that our method greatly improves the sample efficiency of applying RL in IC. Our code is available at https://anonymous.4open.science/r/RLIMFG4IC-811D/

cs.LG↗

arXiv2024

Individual Contributions as Intrinsic Exploration Scaffolds for Multi-agent Reinforcement Learning

In multi-agent reinforcement learning (MARL), effective exploration is critical, especially in sparse reward environments. Although introducing global intrinsic rewards can foster exploration in such settings, it often complicates credit assignment among agents. To address this difficulty, we propose Individual Contributions as intrinsic Exploration Scaffolds (ICES), a novel approach to motivate exploration by assessing each agent's contribution from a global view. In particular, ICES constructs exploration scaffolds with Bayesian surprise, leveraging global transition information during centralized training. These scaffolds, used only in training, help to guide individual agents towards actions that significantly impact the global latent state transitions. Additionally, ICES separates exploration policies from exploitation policies, enabling the former to utilize privileged global information during training. Extensive experiments on cooperative benchmark tasks with sparse rewards, including Google Research Football (GRF) and StarCraft Multi-agent Challenge (SMAC), demonstrate that ICES exhibits superior exploration capabilities compared with baselines. The code is publicly available at https://github.com/LXXXXR/ICES.

cs.LG↗

arXiv2023

Cooperative Operation of the Fleet Operator and Incentive-aware Customers in an On-demand Delivery System: A Bi-level Approach

In this paper, we study the cooperative operation problem between the fleet operator and incentive-aware customers in an on-demand delivery system. Specifically, the fleet operator offers discounts on transportation costs in exchange of the delivery time flexibility of customers. In order to capture the interaction between the fleet operator and customers, a novel bi-level optimization framework is proposed. By exploiting the strong duality, and the KKT optimality condition of customer optimization problems, we can reformulate the bi-level optimization problem as a mixed integer nonlinear programming problem. Considering the inherent difficulties of MINLP, a computationally efficient algorithm, which combines the merits of Lagrangian dual decomposition and Benders decomposition, is devised to solve the resulting MINLP problem in a distributed manner. Finally, extensive numerical experiments demonstrate that the proposed cooperation scheme can decrease the delivery fees for the customers, and reduce the operation cost of the fleet operator at the same time, thus leading to a win-win situation for both sides.

eess.SY↗

arXiv2023

Security Verification of Low-Trust Architectures

Low-trust architectures work on, from the viewpoint of software, always-encrypted data, and significantly reduce the amount of hardware trust to a small software-free enclave component. In this paper, we perform a complete formal verification of a specific low-trust architecture, the Sequestered Encryption (SE) architecture, to show that the design is secure against direct data disclosures and digital side channels for all possible programs. We first define the security requirements of the ISA of SE low-trust architecture. Looking upwards, this ISA serves as an abstraction of the hardware for the software, and is used to show how any program comprising these instructions cannot leak information, including through digital side channels. Looking downwards this ISA is a specification for the hardware, and is used to define the proof obligations for any RTL implementation arising from the ISA-level security requirements. These cover both functional and digital side-channel leakage. Next, we show how these proof obligations can be successfully discharged using commercial formal verification tools. We demonstrate the efficacy of our RTL security verification technique for seven different correct and buggy implementations of the SE architecture.

cs.AR↗

arXiv2022

Joint Routing and Charging Problem of Electric Vehicles with Incentive-aware Customers Considering Spatio-temporal Charging Prices

This paper investigates the scheduling problem of a fleet of electric vehicles, providing mobility as a service to a set of time-specified customers, where the operator needs to solve the routing and charging problem jointly for each EV. Hereby we consider incentive-aware customers and propose that the operator offers monetary incentives to customers in exchange for time flexibility. In this way, the fleet operator can achieve a routing and charging schedule with lower costs, whilst the customers receive monetary compensation for their flexibility. Specifically, we first propose a bi-level optimization model whereby the fleet operator optimizes the routing and charging schedule accounting for the spatio-temporal varying charging price, jointly with a monetary incentive to reimburse the delivery time flexibility experienced by the customers. Concurrently the customers choose their own time flexibility by minimizing their own cost. Second, we cope with the computational burden coming from this nonlinear bi-level optimization model with an accurate reformulation approach consisting of the KKT optimality conditions, a Big-M-based linearization method, and the zero duality gap of convex optimization problems. This way, we convert the proposed problem into a single-level optimization problem, which can be solved by a strengthened generalized Benders decomposition method holding a faster convergence rate than the generalized Benders decomposition method. To evaluate the effectiveness of the proposed mathematical model, we carry out numerous simulation experiments by using the VRP-REP data of Belgium. The numerical results showcase that the proposed mathematical model can reduce the delivery fees for the customers together with the cost of operation incurred by the fleet operator.

math.OC↗

arXiv2022

Incentive-aware Electric Vehicle Routing Problem: a Bi-level Model and a Joint Solution Algorithm

Fixed pickup and delivery times can strongly limit the performance of freight transportation. Against this backdrop, fleet operators can use compensation mechanisms such as monetary incentives to buy delay time from their customers, in order to improve the fleet efficiency and ultimately minimize the costs of operation. To make the most of such an operational model, the fleet activities and the incentives should be jointly optimized accounting for the customers' reactions. Against this backdrop, this paper presents an incentive-aware electric vehicle routing scheme in which the fleet operator actively provides incentives to the customers in exchange of pickup or delivery time flexibility. Specifically, we first devise a bi-level model whereby the fleet operator optimizes the routes and charging schedules of the fleet jointly with an incentive rate to reimburse the delivery delays experienced by the customers. At the same time, the customers choose the admissible delays by minimizing a monetarily-weighted combination of the delays minus the reimbursement offered by the operator. Second, we tackle the complexity resulting from the bi-level and nonlinear problem structure with an equivalent transformation method, reformulating the problem as a single-level optimization problem that can be solved with standard mixed-integer linear programming algorithms. We demonstrate the effectiveness of our framework via extensive numerical experiments using VRP-REP data from Belgium. Our results show that by jointly optimizing routes and incentives subject to the customers' preferences, the operational costs can be reduced by up to 5%, whilst customers can save more than 30% in total delivery fees.

eess.SY↗

arXiv2021

Evacuation Problem Under the Nuclear Leakage Accident

To handle the detrimental effects brought by leakage of radioactive gases at nuclear power station, we propose a bus based evacuation optimization problem. The proposed model incorporates the following four constraints, 1) the maximum dose of radiation per evacuee, 2) the limitation of bus capacity, 3) the number of evacuees at demand node (bus pickup stop), 4) evacuees balance at demand and shelter nodes, which is formulated as a mixed integer nonlinear programming (MINLP) problem. Then, to eliminate the difficulties of choosing a proper M value in Big-M method, a Big-M free method is employed to linearize the nonlinear terms of the MINLP problem. Finally, the resultant mixed integer linear program (MILP) problem is solvable with efficient commercial solvers such as CPLEX or Gurobi, which guarantees the optimal evacuation plan obtained. To evaluate the effectiveness of proposed evacuation model, we test our model on two different scenarios (a random one and a practical scenario). For both scenarios, our model attains executable evacuation plan within given 3600 seconds computation time.

math.OC↗

arXiv2021

Joint Routing and Charging Problem of Multiple Electric Vehicles: A Fast Optimization Algorithm

Logistics has gained great attentions with the prosperous development of commerce, which is often seen as the classic optimal vehicle routing problem. Meanwhile, electric vehicle (EV) has been widely used in logistic fleet to curb the emission of green house gases in recent years. Solving the optimization problem of joint routing and charging of multiple EVs is in a urgent need, whose objective function includes charging time, charging cost, EVs travel time, usage fees of EV and revenue from serving customers. This joint problem is formulated as a mixed integer programming (MIP) problem, which, however, is NP-hard due to integer restrictions and bilinear terms from the coupling between routing and charging decisions. The main contribution of this paper lies at proposing an efficient two stage algorithm that can decompose the original MIP problem into two linear programming (LP) problems, by exploiting the exactness of LP relaxation and eliminating the coupled term. This algorithm can achieve a nearoptimal solution in polynomial time. In addition, another variant algorithm is proposed based on the two stage one, to further improve the quality of solution.

math.OC↗