Model Optimasi Rute Pengangkutan Sampah dengan Cummulative Vehicle Routing Problem (CVRP) Menggunakan Algoritma Genetika dan Prediksi Volume Sampah Dinamis dengan Xgboost Regression

Abstract

Waste collection truck route management is a major challenge in an effort to improve operational efficiency. Cummulative Vehicle Routing Problem (CVRP) is one of the problems in vehicle route optimization, especially in the context of waste transportation that considers vehicle capacity factors, operational time and dynamic waste capacity. This research proposes a Machine Learning-based optimization model to improve the efficiency of CVRP. This model utilizes the prediction of waste volume at each collection point using a regression algorithm to generate dynamic data that is more realistic than the static approach. Furthermore, the prediction results are used in vehicle route optimization by applying Genetic Algorithm (GA). The results showed that the integration of Machine Learning-based prediction with GA optimization was able to improve route efficiency, experimented on datasets with up to 2200 customers and 20 vehicles with reductions ranging from 1.79% to 12.75%. The most significant improvement was seen in East Sidorame, where the optimization distance was reduced by 12.75%, indicating high accuracy in route optimization. The proposed algorithm achieved a 6.46% improvement in solution quality compared to the traditional greedy algorithm. The developed model is also more adaptive to changes in actual conditions, thus providing a more optimal solution compared to conventional methods