In the last years, supplier selection, one of the most used methods on discovering best supplier or suppliers to produce or deliver products based on some restrictions or criteria, became more and more complex, trying to include various elements into model. The manufacturing and supply chain processes have changed significantly, for various reasons, such as technological evolution or COVID-19 pandemic or for a competitive advantage. Various studies showcased the benefits of Industry 5.0 on creating a strong connection between humans and robots, thanks to the Artificial Intelligence and Machine Learning algorithms functions, which replace the repetitive activities, and improves productivity. Since COVID-19 was an unexpected event which affected the entire world, the supply chain process confronts numerous problems when the global economy stopped, and when it started back, the prices increased significantly, dealing with new problems, understanding the need of multiple sources. COVID-19 changed the business’s workflow, by taking into consideration at any moment an external risk which could alter the activity, and making them more cautious, preferring to raise the acquisition costs by diversifying suppliers. Purchasing cost represents a competitive advantage, most of the companies are making a profit based on the acquisition price. Decision-making processes are sensible to numerous components, and it’s mandatory to find an equilibrium between conflicts tangible and intangible elements. Our research analyzes a client with multiple depots, having various suppliers which offer different quantities and price levels. To address this complexity, we explored multiple methods, including Mixed Integer Non-Linear Programming (MINLP), Fuzzy Logic, and the Greedy algorithm. MINLP was created to solve dynamic supplier selection problems, allocating simultaneous quantities to the selected suppliers, being able to be customized, influencing the final price. Greedy algorithm uses fractional knapsack method, being very powerful on solving large scale problems, up to 100.000 suppliers in a few seconds, with high accuracy. The Greedy algorithm was tested in a several methods for batch delivery and supplier selection in a manufacturing domain, and in the majority of cases, the algorithm provided an optimal result, demonstrating the model’s robust fit, and showcasing a high level of effectiveness. Fuzzy logic has been used to convert the preferences expressed in text into fuzzy numbers, calculating fuzzy scores, translating into crisp scores which allows to create a ranking for suppliers, being applied to operations management, psychology, and mathematics.