A network-DEA model with new efficiency measures to incorporate the dynamic effect in production networks

A production network can be described as a collection of production processes performed by several interdependent groups of sub decision-making units (SDMUs) within a DMU. Dynamic effects pertain to the situation where intermediate outputs consumed by one SDMU may also dynamically influence its output level in the future. Without considering these effects in efficiency measurement, we would obtain biased efficiency measurement, because the measure could not faithfully reflect the underlying performance. Hence the result would provide misleading information to decision-makers. This paper proposes a network-DEA model with new efficiency measures to systematically cope with the dynamic effect within a production network. Various interconnections between the new measure and the DEA-efficiency have also been established. Additionally, we also formalize the relationship between returns-to-scale properties of DMUs and those of its constituting SDMUs. This paper presents a unified framework to analyze performances in a dynamic production network.