Optimization approaches for recyclable municipal solid waste collection

Waste management or rather materials management is an important policy subject. Residents, legislators and planners are confronted with ever increasing targets for reused or recycled material flows. In this thesis, the focus is on the design of collection schemes for recyclable municipal solid waste. The toolbox offered by the field of combinatorial optimization is used to find optimal or near-optimal solutions for these waste collection problems. After a thorough review on the literature on strategic network optimization models in waste reverse supply chains, two real-life waste collection problems are considered. The first one is about the collection of recyclables from households on a tactical level. The second one concerns the collection of containers from civic amenity sites and is situated on a tactical-operational level. The models aim at minimizing the total cost, which consists of vehicle and personnel costs. Both problems prove hard to solve. Since a general purpose solver could not find satisfactory results, alternative solution approaches, based on a problem decomposition, are proposed. A mixed integer linear programming based heuristic, combined with a column generation lower bound procedure gives good results for the first problem. For the container collection problem however, two alternative approaches are developed. A first one is a cutting plane heuristic based on combinatorial Benders decomposition. A second one applies a route generation and route selection scheme. At least for the problem instances considered in this thesis, the latter outperforms the former. Therefore, the route generation and selection solution process is embedded in an optimization-simulation model. This model incorporates, next to cost, also environmental impacts in its objective function. Additionally, the solutions are simulated to evaluate their behavior under real-life conditions.

Publication year:2019

Accessibility:Open