Application of Multi-Output Neural Networks for Modeling Sustainable Value Chain Costs and its Impact in Cost Management and Performance Improvement: An Applied Study in State Company for Petrochemical Industries

This research aims to study the application of multi-output neural networks in modeling the costs of the sustainable value chain and to analyze their role in cost management and improving institutional performance, through an applied case study in the State Company for Petrochemical Industries. Industrial enterprises face increasing challenges in balancing economic efficiency with environmental and social considerations, necessitating the use of intelligent analytical tools to support decision-making. In this context, artificial neural networks emerge as an effective tool for estimating the complex nonlinear relationships between value chain variables and institutional performance dimensions. The study adopted an applied analytical methodology, collecting quantitative data from the selected company, including production operation costs, sustainability indicators, and institutional performance outcomes. A multi-output neural network model was constructed using deep learning techniques to predict sustainable costs (economic, environmental, and social) based on a set of operational variables. The study results showed that the proposed model has high predictive accuracy and significantly contributes to enhancing cost management effectiveness and supporting strategic planning towards improving overall institutional performance. Furthermore, the study demonstrated the potential of employing artificial intelligence in reinforcing sustainability concepts within the Iraqi industrial sector. The research recommends that industrial enterprises adopt intelligent analysis and digitization technologies in cost management and integrate sustainability dimensions into management accounting systems to ensure continuity and improve institutional performance in a competitive and dynamic environment.