Enhancing Vendor Selection in Supply Chains Using Machine Learning: A Comparative Study of Optimization Algorithm

Optimal supplier selection step is a vital feature of the supply chain, having a great effect on the company's financial and operating results. The standard methodologies that relied on the lowest rate acceptances have been changed for this purpose as they should be replaced with a more in-depth one that uses multidisciplinary sourcing teams and the most sophisticated analytics. This paper suggests a predictive model for building an appropriate tendering system with machine learning methods. The evaluation model is developed with pre-established criteria that measure the vendor performance on such aspects as the quality, cost, and availability. This generates a combined score useful for successful choices. Data of previous years and optimization algorithms like Nadam, Adam and RMSprop are used and their efficacy is assessed in terms of predicting the model. Neural network architecture developed is already showing the strength of data-driven methods and this is how the management can be better optimised and vendor selection practice improved. The results should have widened the scope of the optimization methods and should emphasized an appropriate selection of optimization algorithms for the high performance.