Customer grouping for better resources allocation using GA based clustering technique

Appropriate organizational resources allocation becomes a major challenge for companies to address the rapid demands for resources from different operational aspects while resource utilization is keeping low. Differentiate exiting customers with common features into smaller groups can serve as a piece of useful reference for decision-making. So far, k-means algorithm is the most commonly used clustering technique for conducting customer grouping. However, k-means limits the grouping consideration to a fixed number of dimensions among each group and the grouping results are significantly influenced by the initial clusters means. In this research, a robust genetic algorithm (GA) based k-means clustering algorithm is proposed in attempt to classify existing customers of the enterprise into groups with consideration of relevant attributes for the sake of obtaining desirable grouping results in an efficient manner. Different from k-means, the proposed GA-based k-means algorithm is able to select which and how many dimensions are better to be considered for each customer group when developing approximate optimal solutions. A case study is conducted on a window curtain manufacturer with the application of software Generator associated with MS Excel.