logistics coordination

Existing conference scheduling tools rely on static visualizations, failing to accommodate participants’ real-time inputs or personalized preferences. This paper proposes a large language model (LLM)-driven dynamic meeting scheduling system that—uniquely—integrates LLMs throughout the entire agenda coordination pipeline. The system achieves dual adaptivity: (1) dynamically generating time-slot recommendation pools via semantic intent understanding and context-aware availability modeling, and (2) producing explainable, human-readable visual interface representations. By automating availability articulation and enhancing transparency, it substantially reduces users’ cognitive load in expressing availability while improving organizers’ decision efficiency and quality. In a controlled experiment with 66 participants, the system significantly outperformed both oral communication and shared-calendar baselines across coordination efficiency and user satisfaction metrics.

Hybrid hackathons face unique challenges—including misalignment between physical and virtual spaces, synchronization breakdowns, and insufficient support for dynamic mode transitions—while existing research lacks systematic integration of organizational design and participant practices. Drawing on hybrid collaboration theory, this study develops a four-dimensional analytical framework encompassing synchronization, spatial distribution, dynamic transition, and technical architecture. Through ethnographic observation, in-depth interviews, and collaborative behavioral data analysis across three hybrid hackathons, we examine how implementation modes (offline-dominant, online-dominant, or balanced) differentially affect resource access, communication efficiency, and team coordination. Key findings reveal critical organizational blind spots in preparatory design; participants’ reliance on adaptive, self-organized strategies to compensate for infrastructural and procedural gaps; and the pivotal role of time management and seamless physical-virtual transition mechanisms in determining collaborative efficacy. The study contributes both a theoretically grounded framework and empirically informed practical guidelines for designing and facilitating hybrid innovation events.

This work proposes a behavior-based planning approach for generating diverse solutions tailored to complex scenarios such as risk assessment, urban design, and game evaluation, where varied planning outcomes are essential. By explicitly modeling diversity within the planning process and supporting multiple objective types, the method enables the production of heterogeneous yet purposeful plans. The study pioneers the application of the behavior planning paradigm across three distinct domains—narrative generation, urban planning, and game critique—and demonstrates its cross-domain adaptability and practical utility through case studies and qualitative evaluations. Beyond advancing the theoretical foundations of diversity-aware planning, this research establishes a reusable framework and evaluation methodology for developing intelligent systems capable of producing diverse, context-sensitive solutions.

Group decision-making scenarios—such as conference scheduling—face challenges including heterogeneous preferences, asymmetric power dynamics, and inefficient interpersonal interaction. Method: This paper proposes a large language model (LLM)-based collaborative decision-making framework. It introduces a novel large-scale synthetic dialogue simulation paradigm for evaluation and implements a two-stage mechanism: (1) heterogeneous preference modeling via dialogue state tracking and employee profiling; and (2) multi-objective iterative optimization balancing fairness and satisfaction. The framework integrates LLM-based reasoning, preference-balanced aggregation, and dynamic solution generation. Contribution/Results: Experiments demonstrate that the framework significantly reduces required interaction rounds while achieving over 92% accuracy in preference aggregation and over 92% plausibility in logical reasoning—both in synthetic simulations and human-subject studies. It establishes a verifiable, LLM-augmented paradigm for collective intelligence in group decision-making.

Novice hackathon organizers frequently encounter failure due to suboptimal modality selection (in-person, online, or hybrid) and operational complexity. Method: Drawing on educational design research, operational practice, and cross-case comparative analysis, this study develops the first unified, modular decision-making framework that systematically integrates physical, digital, and hybrid modalities. The framework centers on 12 core organizational decisions and incorporates a dynamic timeline, resource-constraint analysis, and empirically grounded adaptation guidance. Contribution/Results: Its key innovation lies in deeply coupling all three modalities within a single logical architecture, enabling context-sensitive, on-demand configuration. The resulting open-source, downloadable, and citable planning toolkit has been adopted as a standard preparatory manual by numerous universities and technology organizations worldwide, demonstrably reducing novice organizer failure rates.

This study addresses the challenge of high attrition among new contributors following their initial contribution to open-source projects, where the impact of different onboarding mechanisms on retention remains unclear. Leveraging a matched-cohort design across 330 open-source projects, we compare behavioral patterns and retention outcomes between 2,001 contributors onboarded through organized outreach activities and 2,001 who joined organically. Employing survival analysis and behavioral clustering, we demonstrate for the first time that activity-based onboarding significantly alters participation rhythms, fostering more stable contribution patterns strongly associated with long-term retention. Findings reveal that activity-onboarded contributors exhibit a higher likelihood of becoming core members (12.1% vs. 9.6%), longer median retention duration (8.2 vs. 4.8 months), and evidence of mentor dependency effects.

This work addresses the challenge that presentation authoring is often constrained by factors such as time, audience, and communicative intent, yet existing tools offer little proactive support for leveraging these constraints. Through a qualitative user study, the authors propose the first constraint-driven, multi-session presentation authoring (CMPA) framework, reframing constraints from passive limitations into active design drivers. Based on this framework, they developed ReSlide, a prototype system that enables creators to harness constraints explicitly during narrative construction. User studies demonstrate that ReSlide significantly enhances users’ ability to utilize constraints in shaping their presentations and facilitates flexible content reuse across varying constraint conditions. The findings offer a novel interaction paradigm and design implications for next-generation presentation authoring tools.

This study addresses the significant barriers women in computer science encounter in building professional networks due to intersecting structural factors—including geography, economic status, language, identity, personality traits, and caregiving responsibilities. The work proposes a multidimensional analytical framework grounded in these structural dimensions and employs community-driven qualitative workshops alongside participatory methods to systematically uncover the unique challenges and often-overlooked obstacles women face in network formation. By centering lived experiences, the research not only advances theoretical understanding of gendered networking barriers but also fosters collective reflection and co-design within communities around inclusive practices. Ultimately, it offers both conceptual insights and actionable pathways toward more equitable, accessible, and supportive mechanisms for professional networking in computing.

This work addresses the challenge of reconciling diverse preferences and constraints in group travel planning by proposing the first collaborative planning framework based on multiple large language model (LLM) agents. The framework assigns distinct roles to LLMs and leverages natural language dialogue to facilitate negotiation, while incorporating a configurable workflow orchestrator to enable automatic itinerary generation and real-time monitoring and analysis of the discussion process. Experimental results demonstrate the framework’s effectiveness, uncovering behavioral patterns and decision dynamics among multi-agent negotiations. This approach establishes a novel paradigm for modeling group decision-making processes in complex, preference-sensitive domains such as collective travel planning.

This study investigates the impact of large language models (LLMs) acting as coordinators on consensus formation, fairness in resource allocation, and participant perceptions in real-world group decision-making. Through two incentivized experiments involving 879 participants, we compare real-time text-based discussions among three-person groups tasked with allocating charitable funds, under conditions with and without LLM coordination and across different coordination strategies. Results indicate that while LLM coordination does not significantly enhance group consensus or actual participation equity, it shifts funding allocations toward specific organizations by up to 5.5 percentage points and introduces two key governance risks: “algorithmic steering” and an “illusion of inclusivity.” Participants erroneously perceive discussions as more inclusive and express greater trust in the process, revealing the potential for AI-mediated deliberation to subtly manipulate outcomes and distort subjective experiences.