A sustainable decision-making framework and a mixed-integer formulation for the project portfolio selection problem

• Published in: The Journal of supercomputing 2024-09, Vol.80 (14), p.20743-20792
• Main Authors: Hosseinian, Amir Hossein, Esmaeeli, Hamid
• Format: Article
• Language: English
• Subjects: Compilers; Computer Science; Decision making; Delphi method; Interpreters; Mixed integer; Processor Architectures; Programming Languages; Statistical tests; Sustainable development
• ISSN: 0920-8542; 1573-0484
• DOI: 10.1007/s11227-024-06241-3

This research has concentrated on the project portfolio selection (PPS) in the petroleum industry. In this study, the PPS has been treated as a multi-attribute decision-making (MADM) problem; therefore, a hybrid framework comprising five MADM techniques has been proposed to tackle this problem. Several MADM techniques have been integrated to acquire more reliable decisions and consequently to decrease the risk of failure in the decision-making process. Since the proposed methodology is an MADM-based framework, there was a need to discover the influential attributes on selection of petroleum projects. In this respect, the literature of the PPS has been comprehensively reviewed and the most influential attributes have been detected. Sustainable development has been a concern for the researchers; hence, the sustainability-related attributes have been embraced in the decision-making process as well. To strengthen the practicality of the developed framework, the Delphi method has been employed to gather and converge the viewpoints of experts on the identified attributes. The Kruskal–Wallis statistical test has been implemented to compute the weights of the attributes. Having determined the influential attributes and their weights, the embedded MADM techniques in the proposed framework have been implemented to prioritize the potential petroleum projects of a real case study. To obtain the ultimate ranking of alternatives, the proposed framework consolidates the outputs of the aforementioned techniques through using the Copeland method. This paper has also proposed a mixed-integer mathematical formulation for the PPS problem to assess the precision and validity of the results delivered by the decision-making framework. Comparing the outputs of the proposed framework and the model revealed that the developed framework is capable of providing credible outcomes. Furthermore, several sensitivity analyses have been performed to demonstrate the robustness of the framework.