
ABSTRACT Effective communication is critical for coordinating interdependent decisions in collaborative systems. This study evaluates a structured communication channel that makes partners' reported collaboration beliefs visible alongside payoff‐based thresholds indicating when collaboration is advantageous. 52 participants in 26 pairs completed a collaborative design experiment, with sessions assigned to treatment or control conditions. The study found no statistically significant overall differences in collaboration or efficiency between conditions. However, efficiency declined as differences between partners' structural risk thresholds increased in the control condition, whereas the estimated slope in the treatment condition was close to zero. Evidence for this difference remains tentative. Treatment participants reported collaboration beliefs approximately 20 points higher on a 0–100 scale, and both own and partner reports were positively associated with collaborative choices. When actors chose collaboration despite their own reports favoring the individual option, favorable partner reports were more prevalent in treatment than in control. These findings provide preliminary evidence that structured social and technical information may support coordination when actors' assessments and decision thresholds differ. The study identifies candidate information features for collaborative systems management tools and motivates further investigation.
The integration of Large Language Models (LLMs) into the systems engineering lifecycle presents a transformative opportunity to enhance decision-making, automate documentation, and support complex engineering tasks through intelligent assistance. However, current applications of LLMs in this domain remain fragmented, experimental, and often lack systematic guidance for reuse. This paper presents a structured design pattern catalog for applying LLMs across key phases of the systems engineering lifecycle, including requirements engineering, system architecture, verification and validation, and operations and maintenance. Each design pattern encapsulates a recurring problem, its context, and a proven LLM-enabled solution, supported by real-world scenarios and implementation considerations. By formalizing these patterns, we aim to bridge the gap between state-of-the-art language models and established systems engineering practices, enabling practitioners to adopt LLMs more effectively, consistently, and responsibly. The catalog also addresses cross-cutting concerns such as traceability, human-in-the-loop interaction, and lifecycle integration. A multi-case study approach, incorporating real-world industrial applications, validates the identification and practical utility of these patterns, demonstrating their impact on improving systems engineering outcomes.
Successful Model-Based Systems Engineering (MBSE) includes designing the system architecture and its traceable integration into the development process in order to align system realization. To this end, we propose a framework for reaching end-to-end traceability with a novel degree of automation powered by a cloud application. We explain the framework using the example of a testbench robot and outline its transferability to other systems.
Human-AI collaboration is a defining characteristic of AI-augmented systems and increasingly critical for organizational performance. However, small and medium-sized enterprises (SMEs) lack practical, validated methods to assess how effectively humans and artificial intelligence (AI) function together in their organization and to guide improvements in collaborative maturity. This paper extends and empirically validates the Human-AI Collaboration Maturity Model (HAIC-MM) as a systems engineering assessment framework designed to reflect the operational realities and resource constraints of SMEs. The study introduces a quantitative scoring methodology that incorporates novel balance-based scoring and dependency-aware capability progression to capture interdependencies between human and AI subsystems and their influence on system-level performance and behavior. The framework is operationalized through a web-based platform and evaluated through deployment with ten SME practitioners across diverse industries. Results indicate high practitioner satisfaction with assessment usability, strong alignment between outputs and organizational realities, and clear practical value of the generated reports. The novel balance-based scoring approach effectively revealed misalignments between human and AI capabilities that conventional averaging methods obscure, while dependency logic prevented artificially inflated maturity classifications. Overall, HAIC-MM is shown to be a usable, interpretable, and practically valuable tool for assessing and strengthening human-AI collaboration in SMEs. By empirically validating the model and introducing a novel structured scoring approach, this work advances systems engineering methods for evaluating and managing human-AI collaboration, supporting system integration and lifecycle monitoring of AI-augmented systems.
Evolving objectives and requirements, risk management processes, and external disruptions frequently drive change in engineered systems. Due to these pressures, system architects must anticipate and manage change throughout a system's lifecycle. Yet, managing change is challenging due to the risk of change propagation. Many existing change propagation analysis (CPA) methods provide insight into where change may travel but provide limited understanding of why change propagates. CPA methods also help identify which design elements should be "flexible," yet the meaning of flexibility at the element-level is rarely clarified. We address this gap by defining flexibility of system elements and quantitatively measuring them as local, context-dependent properties. Element flexibility provides designers with insight into the feasibility of modifying system elements, the impact of modifications on the ability to make future changes, and pathways for restoring feasibility after an infeasible modification is made. A high-powered rocket design case is introduced and leveraged to illustrate how tracking element flexibility could address these areas of inquiry in practice. The ability to quantify element flexibility also provides a foundation for future changeability research, serving as a structured approach for determining the specific modifications that propagate change.