Exploratory decision robustness analysis of the theory of constraints focusing process using system dynamics modeling

The classic P and Q problem is used to illustrate Throughput Accounting (TA) concepts proposed by the Theory of Constraints (TOC). Since the release of Throughput Accounting, several studies have been conducted, both to discuss it and to consider uncertainties. However, these uncertainties are included in isolation in each research. This limitation distances the analysis from the reality of companies, in which variations and uncertainties occur simultaneously. This study aims to verify whether decisions based on Throughput Accounting are robust when simultaneous variations occur, bringing the problem closer to the reality of companies. A System Dynamics model based on the P and Q problem was used, including simultaneous variability in different parameters. Finally, an analysis of the prevalent variables for decision-making was performed using Artificial Neural Networks. The results suggest robustness of decisions, even in systems with simultaneous variations. One contribution to the literature consists of the demonstration that the Throughput Accounting heuristic is robust in stochastic conditions, even though the original proposal was deterministic. Another contribution lies in the identification of the prevalent variables, so far unexplored by the literature. Finally, the robustness of Throughput Accounting to support product mix and constraint identification decisions in companies is evident in dynamic scenarios that feature variability.