Sustainable reverse logistics for household plastic waste

Summary of the thesis titled “Sustainable Reverse Logistics for Household Plastic Waste” PhD Candidate: Xiaoyun Bing Recycled plastic can be used in the manufacturing of plastic products to reduce the use of virgin plastics material. The cost of recycled plastics is usually lower than that of virgin plastics. Therefore, it is environmentally and economically beneficial to improve the plastic recycling system to ensure more plastic waste from households is properly collected and processed for recycling. Plastic waste has a complex composition and is polluted, thus requires a substantial technical effort to separate the plastics from the waste and to sort these into recyclable materials. There are several alternatives in the existing collection methods (curb-side and drop-off) and separation methods (source separation and post-separation). It is challenging to select a suitable combination of these methods and to design a network that is efficient and sustainable. It is necessary to build a suitable, efficient and sustainable recycling network from collection to the final processor in order to provide solutions for different future scenarios of plastics household waste recycling. Decision support is needed in order to redesign the plastic waste reverse logistics so that the plastic waste recycling supply chain can be improved towards a more sustainable direction. To improve the efficiency in the recycling of plastic packaging waste, insights are required into this complex system. Insights solely on a municipal level are not sufficient, as the processing and end market are important for a complete network configuration. Therefore, we have investigated the problem at three levels: municipal, regional, and global. Decision support systems are developed based on optimization techniques to explore the power of mathematical modelling to assist in the decision-making process. This thesis investigates plastic waste recycling from a sustainable reverse logistics angle. The aim is to analyse the collection, separation and treatments systems of plastic waste and to propose redesigns for the recycling system using quantitative decision support models. We started this research project by identifying research opportunities. This was done through a practical approach that aimed to find future research opportunities to solve existing problems (Chapter 2). We started from a review of current municipal solid waste recycling practices in various EU countries and identified the characteristics and key issues of waste recycling from waste management and reverse logistics point of view. This is followed by a literature review regarding the applications of operations research. We conclude that waste recycling is a multi-disciplinary problem and that research opportunities can be found by considering different decision levels simultaneously. While analyzing a reverse supply chain for Municipal Solid Waste (MSW) recycling, a holistic view and considering characteristics of different waste types are necessary. Municipal Level In Chapter 3, we aim to redesign the collection routes of household plastic waste and compare the collection options at the municipal level using eco-efficiency as a performance indicator. The collection problem is modeled as a vehicle routing problem. A tabu search heuristic is used to improve the routes. Scenarios are designed according to the collection alternatives with different assumptions in collection method, vehicle type, collection frequency, and collection points, etc. The results show that the source-separation drop-off collection scenario has the best performance for plastic collection, assuming householders take the waste to the drop-off points in a sustainable manner. In Chapter 4, we develop a comprehensive cost estimation model to further analyze the impacts of various taxation alternatives on the collection cost and environmental impact. This model is based on such variables as fixed and variable costs per vehicle, personnel cost, container or bag costs, as well as emission costs (using imaginary carbon taxes). The model can be used for decision support when strategic changes to the collection scheme of municipalities are considered. The model, which considers the characteristics of municipalities, including degree of urbanization and taxation schemes for household waste management, was applied to the Dutch case of post-consumer plastic packaging waste. The results showed that post-separation collection generally has the lowest costs. Curb-side collection in urban municipalities without residual waste collection taxing schemes has the highest cost. These results were supported by the conducted sensitivity analysis, which showed that higher source-separation responses are negatively related to curb-side collection costs. Regional Level Chapter 5 provides decision support for choosing the most suitable combination of separation methods in the Netherlands. Decision support is provided through an optimized reverse logistics network design that makes the overall recycling system more efficient and sustainable, while taking into account the interests of various stakeholders (municipalities, households, etc.). A mixed integer linear programming (MILP) model, which minimizes both transportation cost and environmental impact, is used to design this network. The research follows the approach of a scenario study; the baseline scenario is the current situation and other scenarios are designed with various strategic alternatives. Comparing these scenarios, the results show that the current network settings of the baseline situation is efficient in terms of logistics, but has the potential to adapt to strategic changes, depending on the assumptions regarding availability of the required processing facilities to treat plastic waste. In some of the tested scenarios, a separate collection channel for polyethylene terephthalate (PET) bottles is cost-efficient and saves carbon emission. Although the figures differ depending on the choices in separation method made by municipalities, our modeling results of all the tested scenarios show a reduction in carbon emissions of more than 25 percent compared to the current network. Chapter 6 studies a plastic recycling system from a reverse logistics angle and investigates the potential benefits of a multimodality strategy to the network design of plastic recycling. The aim was to quantify the impact of multimodality in the network in order to provide decision support for the design of more sustainable plastic recycling networks in the future. A MILP model is developed in order to assess different plastic waste collection, treatment, and transportation scenarios. A baseline scenario represents the optimized current situation, while other scenarios allow multimodality options (barge and train) to be applied. With our input parameter settings, results show that transportation costs contribute to approximately 7 percent of the total costs, and multimodality can help reduce transportation costs by almost 20 percent (CO_2-eq emissions included). In our illustrative case with two plastic separation methods, the post-separation channel benefits more from a multimodality strategy than the source-separation channel. This relates to the locations and availability of intermediate facilities and the quantity of waste transported on each route. Global Level After the regional network redesign, Chapter 7 shows a global network redesign. The aim of this chapter was to redesign a reverse supply chain from a global angle based on a case study conducted on household plastic waste distributed from Europe to China. Emissions trading restrictions are set on processing plants in both Europe and China. We used a mixed-integer programming model in the network optimization to decide on location reallocation of intermediate processing plants under such restrictions, with the objective of maximizing total profit under Emission Trading Schemes (ETS). Re-locating facilities globally can help reduce the total cost. Once carefully set, ETS can function well as incentive to control emissions in re-processors. Optimization results show that relocating re-processing centers to China reduces total costs and total transportation emissions. ETS applied to re-processors further helps to reduce emissions from both re-processors and the transportation sector. Carbon caps should be set carefully in order to be effective. These results give an insight in the feasibility of building a global reverse supply chain for household plastic waste recycling and demonstrate the impact of ETS on network design. The results also provide decision support for increasing the synergy between the policy for global shipping of waste material and the demand of recycled material. Conclusions Chapter 8 summarizes the findings from chapters 2 to 7 and provides brief answers to the research questions. Beyond that, the integrated findings combine the results from different decision levels and elaborate the impacts of various system characteristics and external factors on the decision making in order to achieve an improved sustainable performance. Main findings are: Regarding the impact of carbon cost, the results from different chapters are consistent in terms that emission cost is only a small part of the total cost, even when carbon cost is set at its historically highest figure. When carbon price is set to a different value, impact of carbon cost on the change of optimization results is higher on the upstream of the reverse supply chain for plastic waste than the downstream.In Emission Trading scheme (ETS), carbon cap has a larger impact on eco-efficiency performance of the global network than carbon price.On one decision level, models can help to find the ``best option". For example, in the collection phase, the average total collection costs per ton of plastic waste collected for source-separation municipalities are more than twice of the post-separation municipalities' collection costs due to the frequent stops made and idling time at each stop. From the regional network perspective, post-separation scenarios have higher costs and environmental impact than source separation due to the limited number of separation centers compared to the numerous cross-docking sites for source-separation. When combining decision levels, however, it is difficult to find one ``best option" that fits all, as there are contradictory results when looking at the same factor from different decision levels. Through decision support models, we provided clear insights into the trade-offs and helped to quantify the differences and identify key factors to determine the differences.Population density differences in various municipalities influence the performance of curbside collection more than drop-off collection. This information is valuable for decision makers to consider in the decision making process. Finally, managerial insights derived from sustainable reverse logistics for household plastic waste are summarized in conclusion section.