The aims of this study is to improve inventory management at San Pa Tong Agriculture Cooperative Limited’s
Supermarket by determining an appropriate inventory model for purchasing important items in order to reduced inventory
management cost. Inventory items in the supermarket are usually large and it is not practical to manage all items with the same
tight policy. Hence, they need to be categorized into groups based on their important. The traditional method for inventory
classification is the ABC analysis. In ABC analysis, inventory classified into 3 groups based on one criterion, which is the annual
baht usage value. However, one single criteria is not suitable in managing inventory of supermarket. In supermarket, other
criteria such as unit cost, unit price, lead time and inventory turnover also have significant impact on inventory management.
Therefore, multi-criteria inventory classification (MCIC) was applied in this research by using clustering algorithm. Two
clustering algorithms including K-means clustering and X-means clustering were used. Important clusters were selected for tight
control by determining appropriate inventory policy. Inventory management cost before and after improvement for both
traditional ABC and clustering algorithm were calculated. The result revealed that the group clustered by k-means was the most
effective which results in 64.23% or 200,831 Baht reduction in inventory management cost. This method could also be applied to
other supermarkets in order to gain better control of their inventory.