Purpose
The study presents two new mathematical models that will contribute to making the allocation of kindergarten places more efficient and equitable with regard to the child's age, gender, siblings, special educational needs and travel time to kindergarten. The models are based on numerical programming and are used to minimise challenges that may arise in the allocation process, such as when children do not get their preferred kindergarten place or when siblings do not end up in the same kindergarten.
Result
The results show that the tested allocation models work better than the way Tønsberg Municipality currently allocates kindergarten places to children. The models have also been tested on larger cases, and by reducing variables, the researchers claim the models can quickly produce good solutions for any municipality in Norway. The researchers behind the study also believe that the models can help establish a better national standard for how children are allocated kindergarten places. Among other things, more people can get the choices they prefer and driving distances for parents are reduced.
Design
The allocation models were used in a pilot project in Tønsberg Municipality in Norway, involving approximately 2200 children and 43 kindergartens. The solutions from the models were then compared with the solutions from Tønsberg's current allocation system. The models were also tested with larger datasets from the municipalities of Trondheim and Oslo to show that the models can be scaled up to handle even larger numbers of children and kindergartens.
References
Geitle, A. H., Johnsen, Ø. K., Ruud, H. F. E., Fagerholt, K. & Julsvoll, C. A. (2021). "Kindergarten allocation in Norway: An integer programming approach". Journal of the Operational Research Society, 72(7):1664-1673.