Performance assessment of waste sorting: Component-based approach to incorporate quality into data envelopment analysis

Abstract

Waste sorting is an important recovery process in circular product systems. Engineering sciences deal with its performance measurement by analysing various parameter settings based on technical key figures derived from object and component quantities. They are difficult to aggregate, especially since there are no prices for all inputs and outputs. Data Envelopment Analysis (DEA) allows an aggregation without implementing prices or weight restrictions. This paper shows how DEA can be used as performance measurement method for sorting processes, whose purpose it is to separate different material components from a waste mixture as effectively and efficiently as possible. This is achieved when the individual components are sorted out into their respective recycling bins with as little effort as possible and without allowing too many incorrect throws. In this respect, sorting processes represent pure material reallocations – and no material transformations. Thus, valid material balances of sorting technologies do not allow for free disposability. A new multiple criteria DEA approach is suggested, which uses input and output objects that entirely consist of components determining the object's value. Different from common DEA approaches, the technological perspective of modelling the underlying production system is strictly separated from the preference perspective of modelling multiple performance criteria for the relevant input and output types and sorts. In more detail, we develop DEA models for a combined evaluation of operating and quality efficiency where both inputs and outputs are characterised by sorts or grades of varying quality. In contrast to existing approaches that map quality into DEA we reflect upon the underlying technological assumptions and analyse which dominance relations result from that in our case of waste sorting.