Sustainable clustering of customers using capacitive artificial neural networks: a case study in Pegah Distribution Company

¹ SAP, Software Company, Tehran, Iran
² Young Researchers and Elite Club, Karaj Branch, Islamic Azad University, Karaj, Iran
³ Faculty of Business, Sohar University, Sohar, Oman

^* Corresponding author: farzipour@yahoo.com

Received: 16 October 2019
Accepted: 31 May 2020

Abstract

To survive, organizations should inevitably work based on sustainability principles in an ever-increasing changes of markets. Appropriate flexibility and responsiveness are particularly important when considering sustainability issue and market changes in clustering problem. One of the uses of clustering can be allocation of resources and equipment for providing the highest level of customer service which has been a matter of concern for decision makers in distributive companies. Capacitive clustering is a common method for solving allocation and distribution problems. However, traditional clustering models ignore sustainability criteria in defining clusters’ capacity. The objective of this study, therefore, is to propose a novel method for optimizing resource allocation for customers given the sustainability criteria. Capacitive clustering is a technique that has a widespread application in data mining. This approach has been used for equipment distribution, sales targeting, market segmentation, etc. One prevalent clustering method is growing neural gas network (GNGN) technique. GNGN is a neural network with uncontrolled learning. In this paper, for the first time, we utilize GNGN to cluster customers given sustainability criteria. Here, the clusters’ centers are determined and allocated with regard to capacity constraints of the clusters. The obtained results in general can be regarded as an optimized sustainable distribution system in which the number of trucks, distribution routes as well as fuel consumptions and environmental pollutions are minimized. We can also refer to reductions in urban traffic, maintenance costs, staff costs, and decreases in the fatigue of drivers and distributers due to the proximity of supermarkets. An illustrative case study is done to indicate the applicability and remarkable contributions of the suggested clustering approach.