Saeid Jafarzadeh-Ghoushchi - Department of Industrial Engineering, Urmia University of Technology, Urmia, Iran
Sona Hashempour - Department of Industrial Engineering, Urmia University of Technology, Urmia, Iran

Urban solid waste (USW) is the natural result of human activities. USW generation modeling is of major significance in programming and planning solid waste management system. Every year, the municipality spends more than %75 of its budget for collection and transportation of solid waste. Waste disposal is essential and is also very expensive. Due to high fluctuation of the amount of the produced waste in Urmia over time, the use of neural networks is appropriate method to optimize and predict the amount of the produced waste based on non-linear and complex relationships between inputs and outputs. In this study, extra parameters such as number of labor, van and truck (waste collection and transport) were employed to assess their effect in improvement structure of ANN model and training performance of generated model. The monitoring data from summer of 2013 are designed to provide the requirements of training and testing the neural network. Finally, with respect to RMSE and R2, suitable models for optimization and forecasting of solid waste were selected for the study. Results point out that artificial neural network model has more advantages in comparison with traditional methods, in optimizing and predicting the municipal solid waste generation.

Urban solid waste, artificial neural networks, batch back propagation, solid waste managemen