Estimating Municipal Solid Waste Generation: From Traditional Methods to Artificial Neural Networks

Abstract

Accurate quantification and prediction of municipal solid waste generation is a prerequisite for establishing an integrated municipal solid waste (MSW) system. Data about both quantity and quality of generated wastes as well as their temporal distribution are essential for planning the MSW system and capacities of the required facilities and equipment. Effective MSW system planning relies both on data availability and reliability and selection of the most appropriate data-driven method for estimating MSW generation. Traditional methods for estimating the amount of generated solid waste are established mostly on the basis of some elements such as number of population, waste generation coefficient per capita, and social-economic development of the region. Conventional methods for predicting MSW generation: weight volume analysis, material balance analysis, and load count analyses are basic methods for estimating the generated waste. Classical statistical models including regression analyses were mainly developed over the last three decades of the twentieth century to estimate MSW generation for specific region/city. Lately, time series models were developed and considered to be more appropriate for predicting waste generation for up to several decades. Artificial Neural Networks (ANNs), relatively new modeling concepts and tools, were successfully used in waste management problems and MSW generation due to their potential to capture temporal effects from data series and reliable predicting of future MSW generation. This paper presents a brief overview on the methods for estimating MSW generation and demonstrates implementation of ANNs to model and predict monthly waste generation for the City of Skopje.