marketplace operations management

This study addresses the challenges of fragmented coordination and delayed responsiveness in retail supermarket supply chains, which stem from reliance on manual decision-making. To overcome these limitations, the authors propose Flowr, a novel framework that integrates a multi-agent architecture with a human-in-the-loop mechanism, enabling end-to-end automated decision-making. Flowr employs a central reasoning large language model (LLM) to orchestrate multiple domain-specific fine-tuned LLMs, while introducing a supervisable collaboration interface based on the Model Context Protocol (MCP) to ensure scalability and accountability. Empirical validation on a real-world large-scale supermarket chain demonstrates that Flowr significantly reduces manual coordination overhead, improves demand–supply matching accuracy, and supports proactive anomaly resolution. The framework exhibits strong potential for generalization across industries.

本文提出一种混合代理AI框架，通过协调代理解析用户意图并分配任务给专门代理，解决供应链分析中的决策难题，提高效率和成本效益。

This study addresses the trade-off between cost and service quality in multichannel customer service by modeling the entire service process as a gated system. It jointly optimizes decisions across three levels: strategic (channel deployment), tactical (staffing and AI allocation), and operational (real-time scheduling). Leveraging operations research, dynamic modeling, and numerical simulation, the work derives a structured optimal request-handling policy and uncovers a counterintuitive insight: judicious deployment of AI chatbots not only enhances service efficiency but also significantly improves service quality, thereby achieving simultaneous optimization of cost and customer experience.

This work addresses the high cost and limited scalability of manually constructing high-quality data products—such as question-answer pairs and database views—which traditionally rely on domain experts. To overcome these challenges, the authors propose an automated optimization framework based on a multi-agent system, introducing for the first time an agent control center architecture. This architecture continuously identifies user queries, monitors multidimensional quality metrics, and integrates a human-in-the-loop mechanism to ensure observability and iterative refinement of data assets. By maintaining human oversight while automating core optimization processes, the approach substantially reduces manual effort and significantly enhances the relevance, coverage, usability, and trustworthiness of data products.

This paper studies the NP-hard problem of offline allocation of Q units of a product across multiple warehouses to maximize the subsequent online order fulfillment rate in e-commerce. We propose the first tight randomized rounding algorithm for the offline surrogate function, achieving a $(1-(1-1/d)^d)$-approximation ratio; under stochastically independent and time-homogeneous demand, it attains a $(1-(1-1/d)^d)/2$-approximation guarantee for the joint allocation-and-fulfillment problem. Our method integrates integer programming modeling, sample average approximation (SAA), and statistical learning-based generalization analysis. Empirical evaluation on real-order sequences from JD.com demonstrates that the proposed offline strategy significantly outperforms myopic and fluid baselines—and even surpasses simulation-based optimization methods. Moreover, the theoretical approximation bound closely matches empirical performance, validating both the efficacy and practicality of our approach.

This study addresses the breakdown of end-to-end guarantees caused by protocol-agnostic design in decentralized agent economies. We introduce "guarantee closure" as a novel task-relative criterion that decouples the evaluation of receipt soundness and completeness. By constructing a six-phase, seventeen-category attribute taxonomy spanning the full task lifecycle, and integrating formal analysis, controlled workflow execution, and large-scale exhaustive verification, this work systematically repairs failure points between verification and settlement. Through 840 matched executions and over ten thousand test cases, we precisely identify and quantify specific failure modes. Ultimately, this research provides both a theoretical foundation and a practical framework for coordinating mechanisms across heterogeneous systems.

This study addresses the challenge of automating workflows in complex industries—such as logistics, healthcare, and construction—where processes are fragmented across heterogeneous tools and involve multi-party collaboration. The work proposes orchestration as a core abstraction to enable effective automation by dynamically coordinating multi-step tasks, enforcing domain-specific constraints, managing human approvals, and integrating legacy systems. It introduces the novel concept of “orchestration bottlenecks” and develops a theoretical framework that unifies multi-agent systems, workflow modeling, constraint reasoning, and human–AI collaboration, while exposing critical gaps in current multi-agent approaches at the orchestration level. Based on distinct sources of operational friction across domains, the paper advocates for targeted architectural safeguards—such as constraint enforcement or explainability—and phased implementation strategies to provide actionable pathways for automation in complex operational environments.

研究通过E-Commerce Bench评估大语言模型在长期自主商业运营中的表现，涵盖市场研究、谈判等多方面，以最大化年末总资产。

本文提出ERPBench，通过模拟企业资源规划中的竞争市场环境，评估大型语言模型在企业决策中的表现及跨生态适应性。

研究通过语言模型解释问题，确定性策略选择并运行预批准分析程序的方法解决企业分析问题，使用关系操作等确保结果可重现。