A greedy heuristics multiple criteria model for solving multi-landfill site selection and plant propagation algorithm for improving waste collection vehicle routing solutions

• Main Author: Nur Azriati, Mat
• Format: Thesis
• Language: English; English
• Published: 2023
• Subjects: QA273-280 Probabilities. Mathematical statistics
• Online Access: https://etd.uum.edu.my/10935/1/Depositpermission_s900803.pdf; https://etd.uum.edu.my/10935/2/s900803_01.pdf; https://etd.uum.edu.my/10935/

This research focuses on two solid waste management (SWM) problems, particularly the landfill site selection problem (LSSP) and the waste collection vehicle routing problem (WCVRP). Solving LSSP involves the evaluation of multiple criteria to determine the best landfill location. Whereas WCVRP involves, the construction of vehicle routes for collecting waste from customers and discharging their loads at the landfills with the minimum total distance travelled. However, there are two main issues with the existing LSSP and WCVRP models. First, previous models focused only on a single landfill site based on the highest score without considering the operational costs criterion in solving LSSP. Second, the Plant Propagation Algorithm (PPA) has never been considered to solve WCVRP. Thus, this research proposes a greedy heuristics multiple criteria model which includes operational costs for solving multi-LSSP and develops PPA for improving WCVRP solutions. First, the importance levels of LSSP criteria including the operational cost criterion were determined by using a modified analytical hierarchy process. Then, a multiple criteria greedy heuristic model was proposed to construct WCVRP solutions and to find a new landfill site(s) with the minimum total operational costs. Moreover, the WCVRP solutions were improved by using PPA. Both models were tested on a WCVRP benchmark problem and a case-based scenario in Kubang Pasu, Kedah. Five candidate landfill sites were considered. The results revealed that a single landfill site (Candidate 4) was the best solution for the case-based scenario, with a 6.74% reduction in total distance travelled. As for multiple landfills, Candidates 3 and 4 were the best alternative sites. Nonetheless, different study areas generated different outputs based on the area's geographical conditions. The WCVRP solutions using PPA are comparable with other best-known solutions on the benchmark problem in terms of total distance travelled and the number of vehicles used. The proposed models were cost-effective and may facilitate the SWM authorities in identifying suitable locations for new landfill site(s) and waste vehicle routes.