A graph with a thousand edges: rummaging in complex work varieties

Context The increasingly digitalization and technological advances ask for novel perspectives to ensure effective risk and safety management strategies. Socio-technical system (STS) have been advocated as constructs able to achieve a certain goal [1], acknowledging both the symbiotic interactions between technical and human elements. Nowadays, these latter have a dominant informative part, demanding for explicit focus on connectivity and communication aspects. This is where the notion of cyber-socio-technical system (CSTS) can be used to broaden the perspective of systemic analyses. CSTSs are systems where peculiar emphasis is reserved to data-accessing and data-processing activities for and from other socio-technical connected entities [2]. Studying - and even more engineering - resilience in such systems becomes a pressing challenge for these modern systems. On this basis, Patriarca et al. discussed the role of Resilience Engineering in combination with knowledge management, to explore different work varieties, as discussed in the WAx framework (Work-As-X) [3]. The framework is made up of three main elements: (i) the knowledge structure, (ii) the knowledge entities, and (iii) the knowledge dynamics, which are meant to capture diverse knowledge entities [4].   Challenge The WAx framework embraces the idea of knowledge dynamics to map the creation and/or conversion of knowledge between agents in both tacit and explicit knowledge dimensions. Consequently, a knowledge model can be built with the intention to abandon any trivialized representation of a work setting and to empower analysts in gaining larger understanding of STSs and CSTSs inherent complexity. Nonetheless, in practical terms, these knowledge models become puzzling to manage and to maintain, requiring an additional systematic approach to make them actionable. This aspect in particular represents one of the modelling and computational frontiers of resilience engineering, in line with the 10th REA symposium “Resilience at frontiers, frontiers of resilience”. Contribution The WAx framework transfers the study of system properties towards the study of the knowledge linked to them, which becomes a big-data management problem. We believe this latter can be tackled as a knowledge graph modelling challenge: a knowledge graph is a model to organize available data based upon the semantic rules of an ontology. Accordingly, a graph G=(V,E) can be defined as a data structure containing a set of vertices V, and a set of edges E connecting them. Each element within the graph is characterized by a label, that classifies each data with an aspect from the common knowledge basis. It is possible to assign properties to each element of the graph with the intention of specifying data values related to certain graph elements. On this basis, a generic vertex in the graph can be defined as: V_n=(L_n^V,p_(i n)^V ) ,0≤i≤I (1) where V_n represents the n-th vertex in the graph (out of the N vertices), L_n^V is the label to be assigned to the n-th vertex, and p_1n^V,p_2n^V,…,p_In^V are the properties that describes n-th vertex. Edges and properties can be defined in a similar way to generate a systematic representation of the phenomenon under investigation. After the selection of a proper ontology, and the subsequent extraction/classification of data through, e.g., natural language processing algorithm [5], the data from the process under analysis can be converted into a knowledge graph. Implications A set of vertices of the graph will represent the system elements to be marked through the agencies mapped via the WAx framework. These latter are all the elements which can generate, transform, or exchange knowledge. On the other hand, any subgraphs of G may represent specific knowledge entities to allow comparing elements (e.g. Work-As-Imagined, Work-As-Done, Work-As-Observed, etc.) across different agents. The interaction between the knowledge entities (i.e., knowledge dynamics) can be explored moving throughout the relationships (edges) connecting nodes (vertices). Such subgraphs can be retrieved by querying the graph and highlighting paths an agent can access offering an unprecedented systematicity to a RE investigation. They permit pinpointing at differences between different work varieties, strengths, ambiguities and weaknesses in the CSTS operations. This research promotes RE as the discipline the frontier of safety and performance management, and pair it with computational advances to ultimately shorten the distance between its theoretical structure and an actionable proactive safety management. References [1] Walker, G., Come back sociotechnical systems theory, all is forgiven … (2015) Civil Engineering and Environmental Systems, 32, pp. 170-179. DOI: 10.1080/10286608.2015.1024112 [2] Monostori, L., Kádár, B., Bauernhansl, T., Kondoh, S., Kumara, S., Reinhart, G., Sauer, O., Schuh, G., Sihn, W., Ueda, K., Cyber-physical systems in manufacturing (2016) CIRP Annals, 65 (2), pp. 621-641. DOI: 10.1016/j.cirp.2016.06.005 [3] Patriarca, R., Falegnami, A., Costantino, F., Di Gravio, G., De Nicola, A., Villani, M.L., WAx: An integrated conceptual framework for the analysis of cyber-socio-technical systems, (2021) Safety Science, 136, art. no. 105142, DOI: 10.1016/j.ssci.2020.105142 [4] Moppett, I.K., Shorrock, S.T., Working out wrong-side blocks, (2018) Anaesthesia, 73 (4), pp. 407-420. DOI: 10.1111/anae.14165 [5] Ansaldi, S., Agnello, P., Pirone, A., Vallerotonda, M. (2021). Near Miss Archive: A Challenge to Share Knowledge Among Inspectors and Improve Seveso Inspections. Sustainability. 13. 8456. 10.3390/su13158456.