Cost estimation in process-based manufacturing environments presents significant challenges due to the presence of multi-level dependencies, dynamic formulation structures, and variability in raw material sourcing. Unlike discrete manufacturing systems, where products are assembled using predefined and relatively stable components, process manufacturing involves continuous or batch transformations in which raw materials are converted into intermediate and finished products through multiple stages. These transformations introduce hierarchical relationships that must be accurately modeled to ensure reliable cost computation. Traditional costing approaches, including spreadsheet-based calculations and generic enterprise resource planning systems, are often inadequate for such environments. They typically rely on static models and simplified assumptions, failing to capture the complexity of hierarchical dependencies and temporal variations. As a result, organizations frequently encounter issues such as inaccurate cost estimation, lack of traceability, and an inability to perform reliable historical analysis. This paper presents a secure and scalable cost intelligence framework specifically designed for process-based manufacturing systems. The framework models production workflows as hierarchical structures, enabling systematic and accurate propagation of costs from raw materials through intermediate stages to finished products. A key feature is the incorporation of a temporal versioning mechanism that associates each formulation revision with an effective date, enabling time-aware cost computation for both current and historical scenarios. The framework also introduces an adaptive data resolution mechanism that combines multi-level matching strategies with self-learning alias generation to address inconsistencies in raw material representation. To ensure confidentiality, a secure multi-database architecture with an isolated bridge layer and mediator gateway regulates data exchange between user roles. The framework is evaluated on a synthetic dataset reflecting real-world manufacturing conditions, demonstrating improved accuracy, consistency, and scalability compared to traditional costing approaches.
Menezes et al. (Tue,) studied this question.