Artificial neural network-based economic decision support approach in post-earthquake disaster waste management: A case study in Adiyaman province

This study focuses on the management of construction and demolition waste generated in the Central District of Adıyaman following the February 6, 2023 Kahramanmaraş earthquakes. It has been determined that approximately 9 million tons of disaster waste were collected in four different areas in Adıyaman Central, which suffered significant damage from the earthquake. Approximately 7 million tons of this amount are located in the Börgenek region landfill, which is actively used. The remaining 2 million tons of waste is scattered across other dumping sites. Within the scope of the study, the processes of sorting disaster waste at its source and in collection areas were examined; the methods applied in the demolition, transportation, and dumping site selection stages in the Merkez district were evaluated. Calculations based on the proportional distribution and average prices of the approximately 21 million tons of demolition waste generated after the earthquakes revealed that the contribution to the state economy was 15.38 billion TL. Within the scope of this thesis, an artificial neural network (ANN)-based economic decision support approach was also applied to estimate the net economic value. In this context, the net economic value was estimated using waste quantities (tons) and unit prices (TL/ton) as inputs; the model's performance was calculated using the RMSE, MAE, and R² metrics. Model outputs; supported by visual assessments such as actual-predicted comparisons, error distribution, ±10% error band, and Bland–Altman analysis; the relative effects of variables were also examined using permutation feature importance analysis.