Conceptual Framework for Prescriptive Analytics Based on Decision Theory in Smart Factories

Prescriptive analytics plays a vital role in decision making processes in smart factories by leveraging available data to provide actionable insights. Time-sensitive decisions related to maintenance, product quality, processes and production management require cost efficient solutions. However, the lack of a comprehensive conceptual framework poses challenges in implementing effective prescriptive analytics solutions. The framework shall serve as a baseline to describe and differentiate modular prescriptive analytics approaches. In this paper, we introduce a four-step prescriptive analytics conceptual framework, which facilitates the development and deployment of prescriptive models in smart factories. Additionally, we emphasize the existing gap between prescriptive analytics and decision theory in the current state of the art. Decision theory already provides sophisticated human centric approaches that can be transferred to the data driven approaches that are part of prescriptive analytics. The conceptual prescriptive analytics framework is derived from real world use cases in smart factories, decision theory frameworks and requirements from literature. We discuss the key components and considerations for each step and highlight their interdependencies in a smart factory environment. Our main contribution is a conceptual framework which enables practitioners and researchers to develop modular and efficient prescriptive analytics implementations for smart factories. We advise practitioners to focus on creating data driven decisions and using the framework as a reference for implementation.