Abstract As a new production procedure Additive Manufacturing will present a time-effective production system when adopted in distributed 3D printing mode. In this case, the distributed manufacturing leads to different challenges such as control between production sites. Based on the cloud infrastructure usage for distributed production systems, the product reliability handling is vital. Moreover, AM is used to produce safety–critical systems components and this product type defines AM as an interesting attack target. This study presents a new extension of uncertain Business Process Management System (uncertain BPMS) architecture for detecting anomaly using this extension capability. This extension has a new component as event-based anomaly detector, where intrusion detection can take place through an integration of process mining and game theory techniques. The proposed component could operate based on pre-processor, conformance checker, and anomaly detection optimizer modules. These modules can intelligently control the AM process activities between expected behavior and actual behavior using distributed event logs, a hybrid of highly accurate algorithms such as Improved Particle Swarm Optimization (IPSO), firefly, and AdaBoost algorithms inside the game theory modeling approach. In this case, the game theory technique as an optimizer provides optimal selection strategies for the proposed component to detect untrusted behaviors. The results of the new extension execution on a case study and its evaluation using Nash Equilibrium (NE) solution indicate that the proposed anomaly detector component is highly accurate in anomaly detection for AM process activities and can detect more attacks successfully through guidance of the game theory framework in the system.

中文翻译：

---

使用博弈论建模方法对增材制造过程活动进行基于过程挖掘的异常检测

摘要 作为一种新的生产工艺，增材制造在分布式3D打印模式下将呈现出一种具有时间效率的生产系统。在这种情况下，分布式制造会带来不同的挑战，例如生产现场之间的控制。基于分布式生产系统的云基础设施使用情况，产品可靠性处理至关重要。此外，AM 用于生产安全关键系统组件，这种产品类型将 AM 定义为一个有趣的攻击目标。本研究提出了不确定业务流程管理系统（不确定 BPMS）架构的新扩展，用于使用此扩展功能检测异常。这个扩展有一个新组件作为基于事件的异常检测器，入侵检测可以通过过程挖掘和博弈论技术的集成进行。提议的组件可以基于预处理器、一致性检查器和异常检测优化器模块运行。这些模块可以使用分布式事件日志智能地控制预期行为和实际行为之间的 AM 过程活动，这是一种高精度算法的混合，例如改进的粒子群优化 (IPSO)、萤火虫和博弈论建模方法中的 AdaBoost 算法。在这种情况下，作为优化器的博弈论技术为提议的组件提供了最佳选择策略，以检测不受信任的行为。