Iranian Journal of Operations Research

A New Decision- Making Method Based on Shannon Entropy Analysis

Shabnam Mohammadi Ardakani

Effective management, as one of the central pillars of organizational success, hinges on the ability to make informed decisions by appropriately combining and coordinating various elements to achieve desired objectives. In decision-making processes, especially in complex scenarios, decision-makers often rely on a multitude of factors to arrive at the most suitable conclusion. This is particularly true in multi-criteria decision-making (MCDM), where decisions are based on evaluating several criteria rather than a single measure of optimality. The growing body of research over recent decades has delved deeply into MCDM methodologies, yet one fundamental aspect remains: the varying significance of the criteria involved. It is critical to accurately determine the weight or importance of each criterion to ensure optimal decision outcomes.
In this paper, we introduce a novel weighting technique designed to address the challenges of assigning weights in MCDM problems, called the Dispersion-based Weighting Method (DWM). This method builds upon the principles of statistical dispersion and offers an efficient alternative to traditional entropy-based weighting methods. The process involves constructing a criterion matrix, followed by the calculation of the mean, standard deviation, and coefficient of variation for each criterion. The weights are then computed based on these statistical measures, providing a robust and straightforward approach for determining the relative importance of each criterion.
To validate the proposed DWM technique, several numerical examples are presented, demonstrating its practical application and effectiveness. Additionally, we compare the results obtained using DWM with those derived from the well-established Shannon entropy method, which is widely used in MCDM applications. The comparative analysis reveals a strong correlation between the two techniques, while highlighting the advantages of the DWM approach, including:

Reduced computational complexity
No requirement for data normalization
Applicability to both positive and negative data sets

The findings suggest that DWM offers a more accessible, efficient, and versatile alternative to traditional methods, particularly in situations where computational efficiency and handling of diverse data types are crucial.

Presenting a model of factors affecting foreign investment risk: multi-criteria decision making and linear regression with a fuzzy approach (Upstream Oil Industries)

mohsen Eshaghinia

The main objective of the research is to rank the risk factors of foreign investment and present a model of their impact on the risk of foreign investment in upstream oil industries. It is descriptive in nature and method, and in terms of relationships, inferential and correlational. The statistical population of the research includes managers and experts in the oil industry, and the sample size was estimated at 90 people by random sampling method. The data collected with questionnaires were analyzed using SPSS and Matlab software. The results showed that according to the experts of the statistical population, political risk is in the first rank of mportance in creating foreign investment risk. Also, in the fuzzy regression method, the correlation between foreign investment risk factors and foreign investment risk is completely significant, and political risk has the greatest impact on foreign investment risk, and economic, social and non-commercial risks are in the next ranks. By examining the overall fit of the proposed model, it was determined that the appropriate power of fit of the proposed model has been able to determine the relationship between the independent and dependent variables of the research well.

Reduced State Flexibility but Increased Connectivity Variability in Autism Spectrum Disorder: Evidence from Dynamic Functional Connectivity

Sara Motamed

Autism spectrum disorder (ASD) is consistently associated with abnormal functional connectivity; resting-state fMRI data were obtained from the ABIDE dataset. Dynamic functional connectivity (DFC) was obtained in an autism-specific subnetwork consisting of 17 regions identified from previous static connectivity analyses. Time-varying connectivity matrices were estimated using a sliding window approach, and recurrent connectivity states were identified using a hidden Markov model. Dynamic measures included state occupancy rate, mean dwell time, and edge-level connectivity variability. Compared with controls, individuals with ASD showed a significant decrease in the occupancy of highly integrated connectivity states (ASD: 28.6 ± 7.4% vs. control: 36.9 ± 8.1%, p < 0.001) and longer dwell times in poorly integrated connectivity states (ASD: 42.3 ± 10.2 vs. control: 31.7 ± 9.5 s, p = 0.002). In contrast, edge-level connectivity variability was significantly increased in ASD, particularly in default mode-limbic connections. Importantly, increased connectivity variability in the default mode network significantly predicted ADOS total scores (β = 0.41), (p = 0.001). These findings suggest a dissociation between reduced network state flexibility and increased moment-to-moment connectivity variability in autism spectrum disorder (ASD).

Intelligent Growing SOM Based Approach for Modeling Brain Patterns in Autism Spectrum Disorder

Elham Askari

Autism Spectrum Disorder is associated with atypical brain function and altered neural patterns. In this study, an unsupervised and interpretable model based on Growing Self-Organizing Maps is proposed for modeling brain patterns in individuals with ASD. Neuroimaging data from the ABIDE I T₁-weighted structural MRI dataset are utilized, and discriminative features, including wavelet coefficients, entropy measures, intensity histograms, and edge-based descriptors, are extracted from brain images. The GSOM model dynamically adapts its topology to the underlying data distribution, enabling effective structural representation of brain patterns. While the learning process is fully unsupervised, class labels are introduced only in a post-hoc evaluation step to assess model performance. Experimental results demonstrate that the proposed framework achieves a classification accuracy of 94.2%, while providing clearer cluster separation and improved structural modeling compared to fixed-topology self-organizing networks. These findings indicate that GSOM is a promising tool for adaptive brain modeling and structural analysis of ASD beyond conventional classification-oriented approaches.