Product Pairing Selection for Promotion using Partitioning Method

Abstract

Slow-moving product is harmful to the business. The slow-moving products take up space and tie up the company’s capital and leave the company with fewer funds to invest in its business. Several factors can cause this issue. There are several methods ranging from statistics to heuristic methods for a company to identify slow-moving inventory but all of them rely on data. In this paper, a partitioning technique from the graph network is proposed to partition the inventories or products into a few clusters. It can help the company to identify what group does the product belongs to and at the same time suggest to the company which product can be paired up or bundled up together to clear up aging and slow-moving products. The partition technique is proposed, and the algorithm is coded using the Visual C++ programming language. The simulation results show that the proposed method can partition the task graph onto smaller subgraphs. The subgraphs called cluster consists of the nodes or products with similar purchase volume (the strong connection between the two nodes). Implementing the partitioning technique could help the companies or managers select the appropriate product to be paired together when doing the promotion.