In modern solid waste recycling, Material Recovery Facilities (”MRFs”) separate post-consumer waste streams into single material products for use in manufacturing. MRFs face a number of challenges including difficulties with staffing, contamination, and highly variable process infeeds, to the point where the average American MRF only achieves roughly 60% of its design throughput. To address these issues, in this work we describe a novel framework for state representation of transient material flows in MRFs and we provide examples of how our framework can be used to design systems for real-time measurement, state estimation, and control of these systems. We provide results both in simulation and in physical experiment which suggest process throughput increases on the order of 40%. We then discuss the implications of this work for the recycling industry as well as suggest direction for future work.