Development for a stable and effective seating allocation algorithm in hot-desking with matching theory and network

Abstract Hot-desking is a work style where employees are not assigned a fixed desk, but can work at any available desk on any given day. In recent years, when diversity in working styles is being emphasized, this type of office is attracting attention. In a hot-desking, the problem is that the freedom to choose your own seat can lead to a lack of fixed seating and hinder collaboration between workers. In addition, while considering seat preferences and proximity to colleagues can improve work efficiency, it is not likely to lead to an increase in informal communication (accidental interaction). In this study, this study developed a seat assignment algorithm that promotes informal communication while considering workers’ seat and colleague preferences. This study combine worker and seating area preferences using matching theory and evaluate the combinations using the Katz-Bonacich’s centrality of the co-worker network. This study proves that the fixed point method can efficiently search for stable matching solutions that take externalities into account. Specifically, it refers to the possibility of searching for matching solutions in a finite number of steps, considering one-sided optimal externalities. In addition, adjustments were made to the basic statistics of the matching preference order to ensure that the externality evaluation by the network is not under- or overestimated. In numerical experiments, we use data generated based on an actual human relationship network to demonstrate the versatility and effectiveness of the algorithm. The statistical determination that the two networks have different structures also guarantees the generality of the algorithm.