AUTOMATIC WEIGHT CRITERIA FOR SAW-BASED DECISION SUPPORT SYSTEM USING GRADIENT DESCENT

Abstract

Modern organizations always utilize a Decision Support Systems (DSS) to have informed decision-making. The Simple Additive Weighting (SAW) method is a prevalent approach used in DSS for evaluating alternatives based on specific criteria. However, the subjectivity inherent in determining criteria weights in SAW poses a significant challenge. This research introduces a new approach, combining the SAW method with Gradient Descent to automatically determine criteria weights. By doing so, it takes advantage of the similarities between SAW and linear equations. Our research shows that this method produces more accurate and unbiased criteria weights, as confirmed by the Mean Square Error (MSE) analysis. In conclusion, incorporating Gradient Descent into the Decision Support System has the potential to greatly improve its effectiveness by automating the criteria weight determination process in various decision-making scenarios, leading to more accurate and less subjective decision support in organizations.