Utility Function and Multidimensional Data Analysis for a Sustainable Management Model

Abstract

This paper proposes a sustainable and systemic model for the transition from traditional Total Cost Management (TCM) practices toward a broader and more integrated Systemic Value Management (SVM) framework, incorporating Environmental, Social, and Governance (ESG) criteria. Building on previous work within the AICE/ICEC Scientific Commission, the study introduces the concept of a general utility function as a unifying and quantifiable measure capable of integrating economic, financial, social, environmental, and institutional dimensions in project assessment. The proposed approach aims to provide a rational foundation for comparing, ranking, and optimizing projects in complex socio-economic environments where traditional financial indicators alone are insufficient. The methodology combines established project evaluation tools (e.g., NPV, IRR, ROI, DSCR, BEP) with social and environmental indicators into a single multidimensional utility function. Each domain contributes a weighted component, normalized within a [0, 1] range, allowing dynamic updates of weights through experience, expert judgment, or data-driven learning. This aggregation model enables both multi-criteria optimization and sensitivity analysis, revealing how variations in weights or component utilities affect overall project ranking and robustness. In this sense, the proposed model serves not only as an evaluative tool but also as a framework for iterative improvement and adaptive management across multiple domains. The paper further demonstrates how Multidimensional Data Analysis (MDA)—including Principal Component Analysis (PCA), Factor Analysis, Multivariate Regression, and Machine Learning—can
be employed to empirically derive or calibrate the partial utilities and weights that constitute the general utility function. This integration between MDA and utility theory establishes a data-informed pathway for constructing systemic value models that evolve through continuous feedback from real-world projects. An illustrative example, implemented both in textual and spreadsheet formats, shows how the approach can be applied to a simplified two-dimensional case (economic and environmental), complete with normalization, aggregation, and sensitivity testing. Finally, the study discusses a theoretical extension of the model toward multidimensional tensor calculus, acknowledging its conceptual relevance for representing higher-order interdependencies (projects × criteria × domains × time × scenarios), while emphasizing that such complexity
is best reserved for future research. The paper concludes that a gradual, data-supported evolution from TCM to SVM/ESG is both feasible and desirable, leveraging artificial intelligence and multidimensional analytics as enabling tools for sustainable, value-centered project management and decision-making.