A Conceptual Framework for AI-Enabled Digital Twin-Based Production Planning

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Modern production systems operate under conditions of demand volatility, resource constraints, equipment disturbances, and changing operational priorities. Under such conditions, conventional planning methods based on static assumptions are often unable to maintain schedule feasibility and system stability. This paper examines the use of artificial intelligence (AI) within the architecture of a digital twin of a production system to support adaptive planning. The study systematizes the limitations of traditional production planning, clarifies the functional role of the digital twin as a dynamic representation of the physical production system, and demonstrates how AI extends that role from passive monitoring to predictive and prescriptive decision support. A conceptual architecture of an intelligent digital twin is proposed, integrating operational data acquisition, state modelling, forecasting, scenario evaluation, and plan correction mechanisms. Particular attention is given to the use of AI for bottleneck prediction, disruption detection, workload balancing, equipment downtime risk assessment, and the generation of recommendations for schedule adjustment. Practical application scenarios are discussed for order priority changes, equipment failures, productivity deviations, and resource shortages. The paper concludes that embedding AI into the digital twin enables a transition from static planning to an adaptive planning model based on continuous situational awareness and scenario-driven decision making.