Collaborative Research: Mitigating Emergent System Behavior through System Evolvability

这个合作项目的研究目标是发现控制复杂工程系统(CES)进化性的原则，并了解如何在面对不可预见的未来需求时使用服务阶段进化来减少负面的紧急行为。在为社会提供基本利益的同时，复杂系统的设计也具有非同寻常的挑战性，因为它们具有(I)复杂的内部交互、(Ii)较长的开发时间、(Iii)与延长的服务窗口相关联的生命周期以及(Iv)表现出意外的紧急行为。为了应对这些挑战，本研究项目将首先进行实证研究，以了解现有的CES是如何在服务中发展以响应紧急需求的。这些研究将被用来洞察为什么某些CES在进化方面比其他CES更有效。然后将探索和验证不同的可进化性度量和模型，以指导设计师操纵CES的演变。最后，本项目将探讨如何将有关服务阶段演进的信息集成到价值模型中，以及不同的演进策略如何影响整体系统价值。如果成功，这项工作将开始解决最近CES研讨会上反复出现的一个主题，强调控制和利用系统进化性的必要性。该项目的结果将通过识别和编目实践工程师可以访问的有用原则，极大地促进对系统设计和系统通过进化减少紧急行为的趋势之间关系的理解。通过探索表征系统可进化性的度量标准，这项研究将促进来自各种设计工具的信息的综合，以描述系统能力并量化各种进化可能性所提供的影响。将这些见解与价值驱动的设计联系起来，将研究进化性和系统价值之间的关系，同时进一步使用和验证这种设计方法。这项工作的经验教训和贡献将通过出版物、可上网的设计夏令营教学模块、与行业支持的互动以及在会议上传播。最后，通过在南加州大学和杨百翰校区举办暑期设计夏令营，学生将获得系统设计方面的第一手经验，并将不同背景和思维的学生聚集在一起。

The research objective of this collaborative project is to discover the principles governing the evolvability of complex engineered systems (CES) and understand how service-phase evolution can be used to reduce negative emergent behavior in the face of unforeseen future needs. While providing essential benefits to society, complex systems are unusually challenging to design because they have (i) elaborate internal interactions, (ii) long development times, (iii) lifecycles associated with extended service windows, and (iv) exhibit unanticipated emergent behavior. To address these challenges, this research project will begin by conducting empirical studies to understand how existing CES have evolved in-service in response to emergent needs. These studies will be used to draw insights about why certain CES are more effective at evolving than others. Different measures and models of evolvability will be then be explored and validated to guide designers in manipulating the evolution of CES. Finally, this project will explore how information about service-phase evolution can be integrated into a value model, and how different evolution strategies impact overall system value. If successful, this work will begin to address a recurring theme from recent CES workshops highlighting the need to control and capitalize on system evolvability. Results from this project will significantly advance understanding of the relationship between system design and the tendency for systems to mitigate emergent behavior through evolution by identifying and cataloging useful principles that can be accessed by practicing engineers. By exploring metrics to characterize system evolvability, this research will facilitate the synthesis of information from various design tools to describe system capabilities and quantify the impact provided by various evolution possibilities. Linking these insights to value-driven design will study the relationship between evolvability and system value while furthering the use and validation of this design approach. Lessons learned and contributions of this work will be disseminated through publications, internet-available design camp teaching modules, interactions with industry supports, and at conferences. Finally, student involvement through summer design camps at both NCSU and BYU campuses will grant first-hand experience in system design and bring together students of diverse backgrounds and thinking.