Multiobjective waste management optimization strategy coupling life cycle assessment and genetic algorithms: Application to PET bottles

摘要

A mathematical model based on life-cycle assessment (LCA) results is developed to assess the environmental efficiency of the end-of-life management of polyethylene terephthalate (PET) bottles. For this purpose, multiobjective optimization and decision support tools are used to define optimal targets for efficient waste management. The global environmental impacts associated with the treatment of PET bottles from their cradle to their ultimate graves (incineration, landfill, recycling by mechanical, chemical or thermal processes) are computed in function of the flow of bottles in the different valorization paths. They are based on the calculation of the impacts involved in each elementary process with a LCA software tool, using the CML impact assessment method. The model takes into account the fraction λ of PET regenerated into bottles that can be further recycled, the global impacts being the cumulative impacts corresponding to each "end-of-life". A nonlinear model for the bottle waste collection stage is considered, reflecting that the more diffuse the flow of bottles is, the more difficult it is to collect and consequently, the more environmentally impacting. The resulting multiobjective problem is to find the allocation of bottles between valorization paths that minimizes the environmental impacts of bottle end-of-lives. It is solved using a genetic algorithm, and the trade-off between environmental impacts is illustrated through Pareto curves. A decision support tool then determines the best compromise among the set of solutions. The model is applied to the case of France in 2010. The variables that minimize simultaneously abiotic depletion, acidification and global warming potential are determined, in particular the number of recycling loops. The approach can be easily adapted to any specific product like bio-based plastics or organic wastes to find the optimal allocation between valorization paths.