Towards Engineering Complex Socio-Technical Systems Using Network Motifs: A Case Study on Bike-Sharing Systems

Abstract The socio-technical system (STS) is an important topic in Systems Engineering and Design Science. Its performance is not only affected by social aspects but also influenced by various technical factors. To understand the relationships and interactions among different components and subsystems in STS, many studies have been done either at individual agent level or at the system level, yet few studies were conducted at the local structural level in such systems. Motivated by this research gap, we developed an approach to analyzing STS based on the network motif theory. In this study, we apply this approach to three bike-sharing systems (BSS) to validate its feasibility. We focus on studying the size-3 motif, the most basic building block of complex networks, and its correlations to a BSS’s rebalancing performance in three different cities, i.e., NYC, Chicago, and Los Angeles. This paper reaches three conclusions. First, both seasonal and city effects play a significant role in affecting BSS’s network structure and its local motif characteristics. Second, the rebalancing issue, i.e., the imbalance between bike returns and rentals, happened at the local transit level can be different from that observed at the system level, and vice versa. Third, the average geographical distance of size-3 trip motifs follows strong patters correlated to the motif structures as well as the number of directed links in a motif. Compared with previous studies, these insights would be beneficial to guiding system designers in engineering STS, particularly from a bottom-up manner (e.g., using mechanisms or incentives), to achieve desired system-level performance. This study also provides an in-depth understanding of the relations between local system structures and system-level performance in an STS, therefore contributes to both the design theory of complex systems and the BSS research community from a new network motif-based perspective.