Fuzzy Logic Modelling to Predict Residential Solid Waste Generation: A Case Study of Baranagar

Abstract

Accurate prediction of waste generation plays a key role in planning of Municipal Solid Waste Management (MSWM) system. Traditional statistical forecasting models require large continuous crisp data and cannot incorporate new data. Hence, development of advanced prediction method is necessary. Fuzzy logic handles dynamic and linguistic data set successfully on limited database. A mathematical model was developed in this research work for predicting Residential Solid Waste (RSW) generation using fuzzy logic considering Baranagar Municipality (BM), Kolkata, India, as a case study. Household population, annual income and building area of households were considered as independent variables for predicting RSW generation. Triangular- and trapezoidal-shaped membership functions were used to describe the dependent and independent variables, and centroid method was applied for defuzzification. Initially, two input variables were taken at a time to identify the correlation with RSW generation. Finally, all three input variables were considered. Model results were found to follow similar trends with actual RSW generation. Average waste generation per household was obtained between 0.5 and 1.0 kg/day. Statistical analysis shows that household population–annual income model is the best-fitted model. However, building area–annual income model shows least R2 value. Surface diagram represents that building area has not much effect on RSW generation and household population is the dominating variable for waste generation.