merchant operations management

Business process optimization remains challenging due to fragmented methodologies across process mining, predictive process monitoring, and process-aware recommendation—each operating in isolation without a unified theoretical foundation or integration framework. Method: This paper proposes a closed-loop optimization framework that systematically integrates Alpha algorithm/Inductive Miner for process discovery, LSTM/Transformer for runtime prediction, collaborative filtering/graph neural networks for action recommendation, and explainable AI (XAI) for interpretability—enabling automated bottleneck identification, anomaly forecasting, and prescriptive optimization from event logs. Contribution/Results: We establish the first unified conceptual boundary, evolutionary taxonomy, and synergy paradigm across the three domains; construct a comprehensive classification schema covering 120+ studies; clarify application scopes and standardized evaluation benchmarks; and deliver an industrially actionable methodology selection guide with validated deployment pathways.

To address low payment routing success rates and insufficient system resilience under dynamic gateway performance, this paper proposes a control-theoretic adaptive dynamic routing framework. Methodologically, it integrates generalized feedback control, reinforcement learning, and multi-armed bandit techniques to construct a closed-loop routing controller: departing from conventional PID designs, it introduces a novel feedback mechanism that enables gateway scores to adaptively converge toward their true success rates, balancing short-term responsiveness with long-term stability. The key contribution lies in deeply embedding control theory into payment routing decisions, enabling real-time, performance-aware closed-loop operation—comprising performance sensing, feedback-driven adaptation, and policy optimization. Empirical evaluation in an online production environment demonstrates that the proposed framework improves transaction success rate by up to 1.15% over rule-based routing, significantly enhancing system reliability and elasticity.

Process mining often yields an overwhelming number of candidate process models, creating decision paralysis for managers seeking actionable insights. Method: This paper proposes a multi-criteria decision-making (MCDM) evaluation framework that jointly incorporates quantitative metrics (e.g., fitness, precision) and qualitative factors (e.g., organizational culture alignment). It systematically integrates MCDM techniques—including the Analytic Hierarchy Process (AHP)—into process model prioritization for the first time, moving beyond purely technical, performance-driven selection criteria. Contribution/Results: The framework enables structured, interpretable trade-offs between operational performance and strategic objectives. Evaluated in a logistics case study, it significantly improves contextual sensitivity and managerial alignment in model selection, facilitating robust, transparent decision-making under competing goals.

This paper investigates the optimal unwind strategy for stochastic order flows in a central risk book (CRB), aiming to minimize transaction costs arising from price impact and bid–ask spread. The core challenge lies in balancing internal hedging (warehousing) against external market execution (externalization). Methodologically, we formulate a stochastic control problem and derive, for the first time, a semi-analytical optimal policy for general stochastic inflow processes; the explicit solution incorporates a correction term based on predictive information about future order arrivals. Robustness is validated via Monte Carlo simulations and multi-scenario numerical experiments. Key contributions include: (i) identifying order flow autocorrelation as the critical determinant for requiring forward-looking adjustments—only martingale-type flows admit myopic execution; (ii) demonstrating substantial reduction in aggregate transaction costs under the proposed strategy; and (iii) introducing a novel, practice-oriented evaluation metric that bridges theoretical rigor with operational feasibility.

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 paper addresses the incompleteness and insufficient semantic coverage in existing translations from BPMN 2.0 to the AI planning language PDDL. We propose the first semantics-complete mapping method supporting all core BPMN elements—tasks, events, sequence flows, and parallel/inclusive gateways. Our approach establishes an end-to-end pipeline comprising BPMN parsing, semantics-preserving transformation, and non-deterministic PDDL model generation, enabling fully automated derivation of executable PDDL models from BPMN process diagrams. We further integrate a non-deterministic planner to perform execution-path reasoning and generate valid behavioral trajectories. Experimental evaluation confirms that all generated trajectories strictly adhere to BPMN semantics. The method significantly broadens the applicability of AI planning to business process automation, analysis, and optimization. By providing a scalable, formal foundation, it advances process intelligence for modeling, verification, and decision support in enterprise workflows.

Existing evaluation metrics focus solely on final task outcomes or unordered agent routing, failing to capture deviations in the compliance of agent execution trajectories within payment workflows. This work proposes Agentic Success Rate (ASR), a trajectory-level metric that quantifies workflow fidelity by comparing actual and expected agent execution sequences at the transition level, decomposed into transition recall and precision. Leveraging the multi-agent payment system HMASP—integrating LLM-based trajectory analysis, transition-level alignment, prompt engineering, and deterministic routing safeguards—we validate ASR’s effectiveness across 18 large language models and 90,000 task instances. ASR-guided diagnostics yield up to a 93.8-percentage-point improvement in Task Success Rate (TSR) for certain models, with GPT-5.2 achieving perfect ASR, thereby uncovering critical process-skipping issues invisible to conventional metrics.

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.

Small and medium-sized enterprises (SMEs) and freelancers face significant challenges in cash flow management due to data scarcity and limited computational resources, rendering mainstream enterprise-grade financial tools impractical. To address this, we propose a lightweight, two-module forecasting architecture tailored for few-shot learning scenarios: a front-end machine learning model performs binary classification of accounts receivable delay risk, while a back-end modular time-series model—designed to handle incomplete historical records—enables fine-grained cash flow forecasting. The system is deployed as a web application with deep integration into the Cluee financial platform. Evaluated in real-world operational settings, the prototype demonstrates substantial improvements in prediction accuracy and decision responsiveness under sparse-data conditions. This work bridges a critical gap in intelligent financial control for micro- and small-scale entities, both technically and practically.

This study addresses the limitations of existing autonomous business process execution approaches, which predominantly focus on control-flow constraints and struggle to support compliance-aware decision-making under multifaceted requirements involving data-aware and temporal conditions. To overcome this gap, the work introduces a unified multi-perspective framework that formally integrates data and time constraints through a numeric planning-based modeling approach. This enables efficient what-if analysis and optimal continuation recommendations for partially executed processes. Experimental results demonstrate that the proposed method not only ensures regulatory compliance but also exhibits strong scalability, substantially enhancing the effectiveness and practicality of autonomous decision-making in AI-augmented business process management systems.