Predictive analysis of electronic waste for reverse logistics operations: a comparison of improved univariate grey models

Growing rates of innovation and consumer demand resulted in rapid accumulation of waste of electrical and electronic equipment or electronic waste (e-waste). In order to build and sustain green cities, efficient management of e-waste rises as a viable response to this accumulation. Accurate e-waste predictions that municipalities can utilize to build appropriate reverse logistics infrastructures gain significance as collecting, recycling and disposing the e-waste become more complex and unpredictable. In line with its significance, the related literature presents several methodologies focusing on e-waste generation forecasting. Among these methodologies, grey modeling approach has aroused interest due to its ability to present meaningful results with small-sized or limited data. In order to improve the overall success rate of the approach, several grey modeling-based forecasting techniques have been proposed throughout the past years. The performance of these models, however, profoundly leans on the parameters used with no established consensus regarding the suitable criteria for better accuracy. To address this issue and to provide a guideline for academicians and practitioners, this paper presents a comparative analysis of most utilized grey modeling methods in the literature improved by particle swarm optimization. A case study employing e-waste data from Washington State is provided to demonstrate the comparative analysis proposed in the study.