Using queueing models to estimate system capacity

Artikel i vetenskaplig tidskrift, 2017

The true capacity of a production or service system is often difficult to estimate with sufficient accuracy. The two main methodological approaches to capacity measurement described in the literature – measurements based on engineering and physical factors and measurement based on time studies – are both characterized by conceptual as well as practical shortcomings. This paper provides analyses on how capacity measurement can be approached using queueing models, thereby eliminating some major drawbacks of other approaches. For example, the suggested method is completely insensitive to the proportional relationship between design and effective capacity, and it cannot cause any behavioural problems during measurement since it does not require direct observation of the system itself. Particularly appealing is the potential for practical application to many contexts in many regions. The validity of the approach is discussed.