[SMARTWEST-D13 - Transparent Processes for "Smart Collaboration" in the mountain Region]

NODES – Nord Ovest Digitale e Sostenibile TRANSPARENT PROCESSES FOR “SMART COLLABORATION” IN THE MOUNTAIN REGION [SMART WEST] SPOKE 4 – DIGITAL INNOVATION TOWARD SUSTAINABLE MOUNTAIN DELIVERABLE 13 This document is part of the project NODES which has received funding from the MUR – M4C2 1.5 of PNRR funded by the European Union – NextGenerationEU (Grant agreement no. ECS00000036). 2 Contents 1. Introduction………......…………………………………….………………………………………………………….3 1.1 Smart collaboration ................................................................................................................................. 4 1.2 Co-working spaces .................................................................................................................................. 5 1.3 Processes.................................................................................................................................................. 6 1.4 Demonstrators .......................................................................................................................................... 7 2. Execute ........................................................................................................................................................... 8 2.1 Multiagent simulation .............................................................................................................................. 9 2.2 NCS Simulation ....................................................................................................................................... 9 2.2.1 Computation of simulation estimates ............................................................................................. 13 2.2.2 Input parameters and estimate computation ................................................................................... 14 2.2.3 Demonstrative outcomes of simulation .......................................................................................... 18 2.3 Strengths and Limits of the Demonstrator ............................................................................................. 21 3. Modeling....................................................................................................................................................... 22 3.1 Monitor .................................................................................................................................................. 23 3.2 Responsibilities ...................................................................................................................................... 23 3.3 Process Alignment ................................................................................................................................. 25 3.4 Demonstrative Outcomes of Process Alignments and Responsibilities ................................................ 26 3.5 Strengths and Limits of the Demonstrator ............................................................................................. 27 4. Improve......................................................................................................................................................... 27 4.1 The process model ................................................................................................................................. 28 4.2 Model Evaluation .................................................................................................................................. 29 4.3 Analysis by way of Process Mining ...................................................................................................... 30 4.3.2 Analysis Methodology .................................................................................................................... 30 4.4 Process Mining Application: a set of Guidelines .................................................................................. 32 4.5 Strengths and Limits of the Demonstrator ............................................................................................. 33 5. Conclusions .................................................................................................................................................. 33 Bibliography ..................................................................................................................................................... 34 3 1. Introduction The focus of Spoke 4 Smartwest flagship project is on contrasting depopulation in the mountain regions. Since the regeneration of mountain areas and villages requires a complex interplay of economic, social, and environmental factors (Grecchi et al. 2025), we believe that fostering social innovation is a cornerstone of all efforts with this aim and that networks of co-working spaces are instrumental to this aim. By social innovation, we mean “new social practices that aim to meet social needs in a better way than the existing solutions, resulting from - for example - working conditions, education, community development or health” (Schwarz 2010). Figure 1 - Overview of the work Figure 1 provides an overall glance of the main concepts that are involved in the research work. As can be seen, the core instrument that we foresee as fundamental for enabling social innovation in the mountain areas is the development of a set of co-working spaces, to be organized in a network. Such places provide attractive features not only for the residents, whose home might not provide adequate working spaces, but also for foreign workers and companies which might be interested in delocating (for some time) their working activities to an area that provides both adequate services (e.g. office spaces, meeting rooms, internet connection) and the benefits of being out of town, in the quiet of nature. Trivially, a co-working space is a place where meeting others is facilitated by the very nature of the place itself. To be truly attractive, especially in the long term, the location of the co-working space must offer attractive features that generally vary depending on the type of user. For instance, locals might find it attractive to be close to schools. 4 Smart collaborations can be either internal to a company, and involve persons with different competences who collaborate on a project and are displaced to a co-working space to support such collaboration, or external, and involve freelancers or professionals, working for different companies, who join to give birth to new initiatives. In both cases, co-working spaces come at hand. Concerning depopulation, co-working is an incentive for people to live and work in the mountain territory. In particular, it is expected to overcome some of the limitations that affect remote working, like bad impacts on personal life-work balance, lack of motivation, and such like (Risi and Pronzato, 2021). On the other hand, smart collaboration heavily relies on digital artifacts, like shared notes, tasks, and business processes, as well as digital tools, ranging from remote conferencing systems to cloud services and process management systems. These are all fundamental assets and play an extremely important role. Before describing the demonstrators, we introduce the main concepts. 1.1 Smart collaboration Smart collaboration is defined as the integration of individual, specialized expertise of knowledge workers to deliver high-quality, customized solutions on complex issues (Gardner, 2016), in opposition to a more traditional organization of work, where a company is structured vertically into specialized silos. In this deliverable, we extend the term to cover any kind of work, i.e., not necessarily restricted to knowledge workers, and we consider it under the scope of NODES - SMARWEST, that is, the promotion of digital technologies for improving the performance and sustainability of companies that are located in mountain areas. In most cases, smart collaboration relies on forms of remote working (a key focus of this deliverable), and on the technological infrastructure and tools that support it. The use of technology allows smart collaboration to be distributed in different ways: competencies may be distributed between different actors; actors may be geographically distributed, and may work for different institutions or be freelancers; the collaboration may happen in places that are different each time and that may be different from an organization working office; collaborations may vary each time in its implementation and actors involved. As one can see, a smart collaboration has a distributed and dynamic nature by definition. Research shows that smart collaboration delivers significant financial and people-related benefits to organizations, such as higher profit margins, greater customer loyalty, talent attraction and retention, and competitive advantage when specialists collaborate across functional boundaries (Burgens-Pas and Seckler, 2022). Bringing together their distinct expertise and knowledge, specialists may also create entirely new types of services that can attract new clients. Of course, such collaborations require trust among the parties, both in the sense of respect for each other’s competence and in the sense of belief in each other’s integrity. 1 Smart collaboration, in our perspective, concerns the processes that occur inside a company as well as the creation of double networks (Malecki, 2011), that is, the creation of networks that are external to the company and that aim at capturing ideas and profit from innovative initiatives that emerge in the outside world. This is also particularly interesting for individual workers who work as freelancers, who would like to expand their activities beyond their usual channels. 1 Da https://clp.law.harvard.edu/research/research-projects/teamwork-and-collaboration/ Teamwork and Collaboration, Heidi K. Gardner. 5 In the project, we focused on the implications of exploiting the so-called “weak ties” that connect the patrons of co-working spaces. We are interested in smart collaborations that arise spontaneously, that is, without being strictly imposed by a controlling entity (an organization), but rather involve persons because of their competencies and needs. As one can see, smart collaborations are, by their nature, dynamic and emergent. The claim is that positioning coworking spaces in a suitable way in the mountain regions may be a driver to reduce depopulation and foster innovation. 1.2 Co-working spaces Co-working spaces are imposing themselves as a consequence of the transformations the world of work has undergone in the last ten years, with an increasing emphasis on flexibility and collaboration (Bouncken et al, 2020), which has led to the rise of the remote working model and of hybrid work models. Such models mix remote working with traditional in-office activities, reshaping the working landscape and gaining widespread popularity. These spaces, designed for shared use by individual workers or small businesses, have proliferated globally. Many co-working spaces are located in large urban areas, near clients –usually highly skilled ICT professionals, freelancers, and creative-class employees (Marino et al., 2023; Grazian, 2020; Mariotti et al., 2023; Belk, 2014). The scarce opportunities for collaboration and innovation between professionals and local businesses, and the difficulty of finding networking and professional growth opportunities, which are typically associated with small, remote communities, are seen as disadvantages. In order to overcome these disadvantages, we study a different model: co-working spaces are reimagined as interlinked hubs, that are situated in small villages, and designed not only to foster the traditional co-working model -such as shared resources and communal working environmentsbut also to serve as focal points for skill retention, attraction, and sharing, providing an infrastructure that fosters smart collaboration. We call such an organization a Network of Co-working Spaces (NCS). As its name suggests, an NCS is a structure that connects many co-working spaces, linking them according to their geographic proximity, and helping specialized “districts” to emerge. More in detail, an NCS has three main characteristics: ● It helps users to find co-working spaces. ● It is meant to be managed and supported by a community – indeed, we expect the NCS to be managed by local (public) mountain/rural communities or organizations. ● It supports the emergence of structured networks of competence districts. The latter point results from the observation that, by establishing these networks around core regional competencies and thematic areas, NCSs encourage residents to engage deeply with their fields while also attracting external talents, who look for specialized collaborative communities. This structure emphasizes local governance, ensuring that the initiatives align closely with regional development goals and that they are integrated into the local socioeconomic fabric. The co-working hubs are no longer merely seen as places of work, but as active centers that foster smart collaborations and entrepreneurial activity. This model leverages the dynamics of shared economies and resource pooling, potentially leading to the birth of new startups, driving local employment, and contributing to a robust and resilient regional economy. 6 Smart collaborations and remote working are parts of a same vision, as can be understood by considering the so-called pillars of remote working, that are: 1. Flexibility: employees and freelancers work where they want, as long as they meet deadlines and objectives. 2. Technology: technologies are an essential asset that links geographically dispersed teams. Some examples are: online collaboration tools, project management systems, virtual conferencing systems, and instant messaging. 3. Focus on results: the focus is more on achievements than on hours worked. Workers are more autonomous in the organization of the activities, and they are encouraged to be responsible. 4. Well-being and productivity: flexibility provided by remote work helps workers to better balance between their job and the constraints of everyday life. Flexibility allows workers to explore by interacting with the other patrons of a co-working space. The same technology that supports team collaboration supports the implementation of new collaborations. Focus on results in this case is the driver: in the case of smart collaborations the focus will be on new or better results. Well-being and productivity come as a by-product because, at its core, the organizational model is that of hybrid work. 1.3 Processes The term business process (or simply process, in this document) is used to identify a set of activities that are performed by a group of individuals in coordination in an organizational and technical environment (Dumas et al, 2018). In certain cases, processes are well-formalized and workers are required to strictly adhere to them, but more often is the case, especially with micro, small, and medium enterprises, that processes exist as best practices or as habits, and are part of the company lore. Thus, they cannot be easily shared or analysed, for instance, to implement improvements. The activities that are part of a business process jointly aim at realizing a business goal. A business process is enacted within a company, but the company may interact with other companies by playing roles in some business processes of other organizations (consider, for instance, a supply chain where a company acts as a client of another that acts as a supplier to the former). Thus, the main features of a business process are: 1) it serves as a coordination medium between actors of the same organization and, potentially, with external collaborators (e.g., customers, suppliers, contractors, freelancers); 2) a process is designed to achieve a business goal (e.g., order management, loan application evaluation); and 3) a process is specified from the point of view of a single organization that is expected to enact it, and that has the control on the way its internal activities are performed, for instance, to meet certain standards. However, a business process may involve other organizations playing specific roles (e.g., think of a supply chain, with suppliers being external companies. In this case, we will talk about cross-organizational business processes. Coordination with external parties is realized through communications (e.g., document/message exchange), often mediated by technology (e.g., messaging, email, forms, IOT solutions). 7 Interestingly, business processes can be designed and their executions can be monitored and analysed automatically in order to understand their adequacy and their effectiveness. Many tools on the market allow companies to support their activities by exploiting the power of business processes. In the context of smart collaborations, business processes must be handled in a different way with respect to what is done traditionally. This is due to the need to take into account aspects that are more typical of remote working, where control over the worker is softer and execution is more oriented towards goal achievement. We relate, nevertheless, our work to the standard business process lifecycle that is used for managing processes inside a traditional organization, underlining the differences between the work that was carried out in the project and the traditional approach. 1.4 Demonstrators This deliverable provides a set of demonstrators of tools that can be applied to foster smart collaboration: 1. NCS impact: the demonstrator 2 consists of a repast-simphony (Repast-simphony) agentbased simulation that takes as inputs population distributions in geographic areas, and geographic maps and computes, based on a set of formulas, the impact of the collocation of a set of co-working spaces in terms of likelihood of new collaborations, change in population number, change of pollution levels due to reduced car traffic. The simulation, whose code can be found here https://github.com/Andrea-s99/Coworking_Building_ Simulation.git, uses high-quality geographic data about the Aosta Valley. The demonstrator’s aim is to prove the feasibility of decision-support systems that can help investors, decision and policy makers in estimating the socio-economic return of possible investments. It also shows the potential of existing tools that can combine geographic data and resources in order to tailor simulations to specific areas, taking into account actual locations and roads. 2. model of co-working spaces: the aim of the demonstrator is to show the application of game theory to address the complexity of co-working networks, as tools/environments that create value by helping people who use them to create new collaborations, and thus increase the range of projects they are involved in. 3. Conformance checking extended with responsibilities. The demonstrator shows how classical techniques for conformance checking can be extended to account for additional information. The demonstrator shows how taking into account information in the form of responsibilities supports a better interpretation of the comparison of real execution with process models. 4. Process mining to support process analysis. The ultimate goal of process mining and business process management techniques is to analyse and improve the process and its representation in a company. We demonstrate the use of these techniques presenting the results of a collaboration of an IT company. We also present a set of guidelines that we developed to replicate the analysis process. 2 The simulator was developed by Andrea Saraceni, in his thesis work, presented on April 1st, 2025. 8 We present the demonstrators by tying them, as much as possible, to the standard five phases of business process management (Dumas et al, 2018), namely: 1) design, 2) modelling, 3) execution, 4) monitoring, 5) improve. Such phases are arranged in a lifecycle (see Figure 2) that accounts for the ideal management of the processes by which a company works and that should undergo constant revisions, e.g., to improve performance, to account for new cases, and to guarantee compliance with regulations. Figure 2. Business process Lifecycle. 2. Execute In order to decide whether to create a new co-working space or a new network of co-working spaces, it is important to foresee several aspects that concern the impact it may have both on the environment in which the new space will be introduced and on smart collaborations. Among these, the impact on the transport infrastructure of people moving from/to the place, CO2 emissions, and such like. Also, the impact on pre-existing co-working spaces could in principle be estimated. To do this, we describe a simulation tool that was developed during NODES. Simulation is a wellknown technique that is also part of the traditional approach of process management. It consists of the identification of several KPIs of interest (such as idle time, costs, and such like) on the definition of a process, and on parameters such as the rate of new instance creation, the average time for activity executions, and so on. Changes in the parameters have an impact on the KPIs. The variations can be assessed using simulations, which support the process of decision making. Specifically, we show how the Repast Simphony tool (repast-simphony) can be configured and used for simulating a Network of Co-working Spaces (NCS), and analyze its impact on smart collaborations, on the geographical area, and on the workers’ satisfaction. 9 2.1 Multiagent simulation Repast Simphony, Recursive Porous Agent Simulation Toolkit (repast-simphony), is a toolkit that serves various purposes: it is an agent-based social simulation toolkit (ABSS), a modeling toolkit, and a cross-platform Java-based modeling system. Repast Simphony has been successfully applied in various application domains, including social sciences, consumer products, supply chains, manufacturing, epidemiology, biomedical systems modeling, and ancient pedestrian traffic. Agent-Based Social Simulation (ABSS) represents the elements of a social system through artificial agents of different scales, placing them within a computer-simulated society to observe their behavior. By analyzing the data, researchers can verify hypotheses and understand emergent behavior, in order to apply these insights to real-world social systems. ABSS integrates three key fields: agent-based computing, social sciences, and computer simulation. Agent-based computing involves designing models and agents, while computer simulation focuses on executing the simulations and analyzing the results. Social sciences contribute theories and concepts that shape the model, enabling the study of social phenomena. An agent is an abstraction that may represent any system, taking decisions and showing a behavior. It is situated in an environment that it can perceive by means of sensors and on which it can act by means of actuators. Concretely, in Repast agents are implemented as Java classes including variables and properties. The latter define the agent’s behaviour, in terms of frequency and priority values. Such values are used by the Scheduler whenever it needs to determine which agent and which method to execute. Icon Description Coworking Building Workers’ Office Coworker moving Coworker at work Service Table 1. Icons that are associated with the elements of an NCS simulation. 2.2 NCS Simulation The simulation was realized by Andrea Saraceni (Saraceni, 2025). It includes various elements, that are graphically represented by the icons in Table 1. Workers are implemented as agents, whose decisions are affected by the presence/absence of co-working spaces. In order to design a simulation, it is necessary to define the factors that need to be computed by the simulation run. In the case of NCS several indicators could be thought of. We chose the following ones: 1. Reduction of worker commutes. 16 Concerning the coworker satisfaction, we compute it as follows: 𝑆 =(1− 𝑇𝑡 𝑇𝑤∗60)∙100 where: ● 𝑆 is the satisfaction index. ● 𝑇𝑡 is the travel time for a coworker (in minutes) ● 𝑇𝑤 is the number of worked hours for a coworker (in minutes). The value 𝑇𝑤 is computed based on a normal distribution and can assume a value in the range 4 - 8 hours. One can observe that coworking (in the nearest co-working building) can reduce the travelling time, thus increasing the satisfaction index. Similarly, it is possible to compute the CO2 emission and the economic cost for workers. Specifically, distance is measured along the route that workers follow to reach their destination, taking into account the available roads along the valleys. The number of kilometers, multiplied by the emission rate per kilometer, produces the overall CO2 emission per each journey. In the simulation, when a worker’s position is less than predefined threshold, the work is considered as “arrived”. Otherwise, he/she will continue to move towards the destination, adapting the position by way of small random variations to obtain a more realistic simulation. The economic cost paid by the worker along the itinerary depends on the distance travelled, on the kind of vehicle and its efficiency, besides exogenous factors like traffic. In general, however, the longer the distance the higher the cost. when computing such factors, we made a simplification consisting in using average fuel consumption. This and distance allow computing the vehicle fuel consumption, and the fuel cost allows then to compute the economic cost for the worker. The formulas that are used to calculate CO2 emissions and fuel costs are: 1. 𝐸𝑙𝐶𝑂₂ = 𝐶𝑂2 𝑚𝑖𝑛 + (𝑟𝑎𝑛𝑑𝑜𝑚∙ 𝐶𝑂2 𝑚𝑎𝑥 − 𝐶𝑂2 𝑚𝑖𝑛 ) 2. 𝐹𝑢=𝑑∙𝐹𝑐 3. 𝐶𝑓𝑢𝑒𝑙 =𝐹𝑢∙𝐶𝑙𝑓 4. 𝐸𝐶𝑂₂=𝐹𝑢∙𝐸𝑙𝐶𝑂₂ 5. 𝐶𝑓𝑢𝑒𝑙,𝑡𝑜𝑡 =∑𝐶𝑓𝑢𝑒𝑙,𝑖 𝑛 𝑖 = 1 ∙𝑆𝑝𝑖 6. 𝐸𝐶𝑂₂,𝑡𝑜𝑡 =∑𝐸𝐶𝑂₂,𝑖 𝑛 𝑖 = 1 ∙𝑆𝑝𝑖 where: ● 𝑛 is the number of ticks used by the simulation, 17 ● 𝑟𝑎𝑛𝑑𝑜𝑚 ∼U(0,1) is a random variable with uniform distribution between 0 and 1, and allows us to determine a random quantity in the specified range, ● 𝐶𝑂2 𝑚𝑖𝑛 is the minimum value for CO2 emissions per liter of fuel and it is a model input parameter, ● 𝐶𝑂2 𝑚𝑎𝑥 is the maximum value for CO2 emissions per liter of fuel and it is a model input parameter, ● 𝑑 is the distance, ● 𝐹𝑐 is fuel consumption in liters per kilometer, ● 𝐹𝑢 is used fuel, ● 𝐶𝑙𝑓 is the cost of fuel per liter, it is an input parameter of the model, ● 𝐶𝑓𝑢𝑒𝑙 is the cost of fuel per single itinerary, ● 𝐸𝑙𝐶𝑂₂ is CO2 emission per liter of fuel, ● 𝐸𝐶𝑂₂ is the CO₂ emission per single itinerary, ● 𝑆𝑝𝑖 is a flag that indicates whether a worker is moving in the considered moment: value is 0 if the worker is working, 1 if the worker is along his/her itinerary, ● 𝐶𝑓𝑢𝑒𝑙,𝑡𝑜𝑡 is the total cost of fuel, ● 𝐸𝐶𝑂₂,𝑡𝑜𝑡 is the total emission of CO₂. The first formula calculates CO2 emission by adding, to the minimal emission, a random variation. Other formulas are trivial, and allow calculating CO2 emissions and economic cost for fuel. Clearly, a worker who can rely on a closeby co-working space will do shorter itineraries and spend less for fuel. CO2 emissions will also be reduced. Earning for local services is computed as follows. ℎ = 𝑋 𝐸𝑡0 = 0 𝐸𝑡𝑖+1 = ∑ 𝐸𝑡𝑖 𝑛 𝑖 = 1 +(ℎ∙0.50)+∑ 𝐸𝑡𝑖 𝑛 𝑖 = 1 ∙𝑘∙𝑟𝑎𝑡𝑒1.5 where: ● 𝑛 is the number of ticks of the simulation, ● 𝑋∼U(4,8) is the working time, in hours, of a worker. It is a random variable with uniform distribution ranging between 4 and 8 hours, 18 ● 𝐸𝑡𝑖 is the actual earning at time i, ● 𝐸𝑡𝑖+1 is the total earning at time i+1, ● ℎ is the total number of working hours, ● 𝑘 is a proportionality coefficient, set to (3∙10−7), ● 𝑟𝑎𝑡𝑒 is the estimate of earning growing, it is a parameter provided in input to the simulation model. Earning depends on local services previous earning, on the total number of working hours by coworkers, and on a proportional coefficient that varies based on the parameter rate and represents an earning increment parameter. The formula is dynamic and different simulations will generally reflect the specific conditions of the run. Rate can be considered the estimate of the percentage of earning increase depending on the presence and the permanence of coworkers. Concerning smart collaboration, we defined the following formula 𝐶𝑖+1 =∑ 𝐶𝑖 𝑛 𝑖 = 1 + 𝐼(𝑁 >=2 ∧ 𝑋 ≤ 𝑃𝑐) where: ● 𝐶𝑖+1 is the number of collaborations at time 𝑖. ● 𝑛 is the time simulation value. ● 𝑃𝑐 is the probability of collaboration (0.8 = 80%), ● 𝑋∼U(0,1) is a random variable following a uniform distribution in the interval [0,1], ● 𝑁 is the number of coworkers. ● 𝐼(𝑋 ≤ 𝑃𝑐) is a function returning 1 if the number of coworkers is greater than or equal to 2, and if the condition 𝑋 ≤ 𝑃𝑐 is satisfied, returns 0 otherwise. ● 𝐶𝑖 is the number of collaborations at time 𝑖. Co-working spaces foster the creation of new professional and personal connections, creating a dynamic ecosystem where ideas, needs, competence, and resources can be shared, promoting innovation and professional growth. Collaborations also contribute to the creation of solid social networks, which improve the sense of community and well-being. 2.2.3 Demonstrative outcomes of simulation Baseline - No Coworking. As a first scenario for the simulation, we consider the case in which coworking is not allowed because the necessary infrastructures are missing. In this case we can only measure the reduction of population. Parameters are set as: - Rate Percentage: 10 19 - Time simulation: 10000 By running the simulation in a context where no co-working spaces are available, we observe that the average number of inhabitants in the considered valleys slightly decreases, passing from 1450 to 1350. We observe that part of the population tends to move to Aosta or outside the region. Local services earnings will consequently slightly decrease, while workers’ satisfaction, CO2 emissions, etc. will not significantly change (reductions being directly associated with population reduction). Population reduction in remote areas depends on many intertwined factors, among which reduction of traditional activities (like agriculture and rearing), lack of opportunities, like qualified work positions, lack of services/specialists like schools and medical doctors, lack of advanced infrastructures that, in turn, does not allow remote work making, the valleys, on the whole, less attractive. Our simulation does not model the complex dynamics of population reduction per se. The realization of tools like the simulator described in these pages is a promising direction for providing decision makers the means for foreseeing the impact of the creation of networks of coworking spaces in alpine valleys. The estimates that were used in the simulation are not related to specific socio-economic models, whose complexity goes beyond the aims of the demonstrator. So, for instance, workers’ satisfaction is simply calculated on the basis of the only measures that are computed in the simulation, which are the distance and the economic cost of fuel. In the real world many other factors are involved. Similarly the earnings of the local services are simply associated with the number of people in a valley while actual dynamics would be more complex and variegated. However, the computation concerning the estimate of the number of smart collaborations is, to the best of our knowledge, a first attempt to capture such a number. We believe that the digitalization of services, by improving access to important resources, reduces the disadvantages of living in a remote area, thus attracting young people more. The improvement of life conditions and the creation of new opportunities fosters local entrepreneurship. 20 Figure 7. The simulation computed features over time shown by means of charts. 21 2.3 Strengths and Limits of the Demonstrator The demonstrator reaches the goal of showing the feasibility of the realization of tools that help foreseeing the impact of a co-working building in a specific geographic area, taking into account information about the population distribution, the characteristics of the population, the local services that are actually present in an area, and also the infrastructures. Of course, the production of an actual tool would require resources that were not available to this exploration, namely: 1. accurate socio-economic and environment-related models; 2. accurate information sources (e.g. linked open data, statistics, demographics); 3. a design and development team. All the formulas and assumptions made in the demonstrator were made for the sake of showing the feasibility and the power of such a kind of tool. So, for instance, the percentages of population that are involved (and how) are nothing more than an educated guess. The vehicles’ fuel consumption is an estimated average. We used geographic information about actual towns and roads but not of actual local services. We considered roads as infrastructure but not, for instance, high speed internet availability. The development of actual systems requires a dedicated project. Nevertheless, in the following section we briefly show how mathematical tools, in particular game theory, can come in handy for modeling purposes. Specifically, we will briefly introduce a model that captures the added value (due to smart collaboration) of an NCS. 22 3. Modeling We briefly show how it is possible to use mathematical tools to study the catalyst effect of an NCS for what concerns smart collaboration, as well as community well-being and local economy. The model is explained in detail in an article that was published by researchers in NODES (Alderighi et al., 2024). The model is targeted to identify ways in which it is possible to maximize the added value, that is generated by the collaboration itself, by exploiting game theory. In the proposal, the NCS is modeled as a game and the target is to maximize the NCS value. The agents are seen as nodes of the network and they can join/create coalitions to collaborate on specific projects or share resources and knowledge. Each coalition has a weight that, in actual settings, can be determined by features like agents having complementary skills, resources that are shared, networking opportunities. We denote by 𝑁 the number of agents. Differently than in the real world this number will not change along the runs. We denote by 𝑆1 and 𝑆2 two sets of coalitions. The first contains cooperation coalitions, that is coalitions where agents share a goal, while the latter contains collaboration coalitions, that is coalitions where an agent contributes to another agent’s goals. Let 𝑉(𝑆) denote the weight of the collaboration/cooperation 𝑆. The co-working game is defined by the characteristic function 𝑉:2𝑛 →ℜ. In particular, we have that: 𝑉(𝑆1) +𝑉(𝑠2) ≤ 𝑉(𝑆1 ∪ 𝑆2). The function is superadditive. Without getting into the details, it turns out that collaborating in an NCS is more advantageous than acting individually (i.e. the value generated by the collaboration is greater than the sum of the values generated by the individual actions of the agents of the NCS). Each agent is an operational micro-reality in the NCS. He or she will have an own role and power in the NCS, depending on specific features (e.g. industrial sector, economic-financial weight, kind/level of innovation/technology). In order to understand the power that can potentially be generated by a system that creates an agent network, it is possible to rely on a Shapley Value. Such value helps to evaluate the contribution of each individual in the formation of coalitions. We denote the Shapley Value of agent 𝑖 as 𝜙𝑖(𝑉𝑎𝑙) , where 𝑉𝑎𝑙 denotes the cumulative value of the agents in a given coalition 𝑆. This value quantifies the specific role of each agent in the network and its impact on the overall collaboration. An agent with a high Shapley Value is particularly important for the network, as it makes a significant contribution to collaboration and cooperation. Once the value of each node has been estimated, it is possible to calculate the total value of the NCS by aggregating the Shapley Values of each node: 𝜙𝑁𝐶𝑆(𝑉𝑎𝑙)=𝛴𝑖=1𝑛 𝜙 𝑆𝑗𝑖(.) where (.) can include parameters related to the characteristic function $V(S)$ that reflect the objectives of the NCS, such as promoting innovation, sharing resources, networking, coalition success and integration of different skills. This allows the calculation to be adapted to the unique characteristics of the network and its nodes. This suggests that the agents involved should be encouraged to cooperate and adopt collaborative strategies to maximize the benefits for all participants. The Shapley Value can be 23 helpful in better understanding each agent's role. In conclusion, the model provides a theoretical framework for understanding and promoting collaboration and cooperation within an NCS, highlighting the importance of creating synergies and added value through collaboration between different agents. Also in this case the mathematical framework is a showcase of the possibilities that are opened by the adoption of formal models on the proof of properties of decision-support systems. An improved model, characterized by the attempt of making a twofold optimization of both the value the NCS can produce and the quality of life of its patrons is currently under development. 3.1 Monitor The monitoring phase of business process management aims at implementing all the necessary measures to understand whether the real executions of the business processes differ from the designed process model. It may seem strange that a process is performed in reality differently from how it is expected to be executed in theory. However, one should keep in mind that the experts of the process are the workers, more than the managers or the designers, and that the actual environment in which processes are executed may or may not support the expected actions, may or may not offer unforeseen alternatives, may or may not present unforeseen cases. This is particularly true in case of smart collaborations, where individuals who do not have a common background nor share consolidated practices, join forces to pursue a common goal. Smart collaborations, indeed, have an horizontal nature, spanning through many roles and functions, while most business processes are pretty specialized on the activities they organize. Therefore, it is often the case that workers/collaborators implement shortcuts or workarounds in order to speed up the process or to handle situations that the process model does not foresee. Identifying these differences may support managers in improving the process by implementing actual usage practices as part of the model and, consequently, in modifying the information systems supporting the processes. Additionally, by identifying discrepancies, managers can take countermeasures where the implemented process is not in line with the law or with the internal policies. On the other hand, also co-working spaces and NCS need to be managed, and their management passes through business processes and can be subject to the same kind of variations, which is equally important to identify. This section provides a demonstration of a technique called conformance checking, aiming at comparing an actual execution with a process model. We enhanced this technique by modeling and accounting for responsibilities on the way the process is executed. Responsibilities, indeed, are an important part of an organization's management. However, in most of the techniques, they are neglected. We demonstrate how taking them into account improves conformance results. Details can be found in (Baldoni et al, 2024). 3.2 Responsibilities Responsibilities are often used for capturing task distribution with the expectation that the one who is responsible for a task will perform it at the right moment and with the right means. This is, for instance, the case of RACI matrices, which are indeed used to specify the different kinds of responsibilities that roles have on activities. As another example, in BPMN the specification of the role that is responsible for the execution of a process activity is modeled by means of lanes. 24 Responsibilities can, however, be more complex. For instance, an activity might be adequately executed only if certain other activities occur, conforming to a given pattern (think of two activities to be performed in sequence because the former provides an input needed by the latter). More in detail, we specify responsibilities declaratively, allowing one to complement a procedural process model with explicit relationships between activities that are not already captured by the process model itself. Such relationships include, but are not limited to, causal relationships, where an activity produces an outcome that is needed by another. So, for instance, in the case of a sequence of activities, the process model does not necessarily express possible causal dependencies. As a consequence, in order to match an execution with a model, classical approaches find the model paths that match the highest number of activities of the execution sequence. Thus, the case in which neither of two consecutive activities is performed always counts as two mismatches with respect to the model. However, the absence of the second activity might be justified by the lack of proper input given the absence of the first activity. We assume that the actors, playing roles in a business organization, do act to support the organization in reaching its goals, and hence they behave in accordance with their responsibilities. Under this assumption, we add responsibilities to the model description and we define a cost of alignment which depends both on the number of mismatches, and on respect of the responsibilities that are defined in the model. While neglected responsibilities always amount to costs, mismatches increase the cost only when they are not consistent with some responsibilities. Therefore, responsibilities not only enable a more informed search for an optimal alignment than just the control flow, but also represent a valuable source of information in the aftermath. In fact, the possibility of reasoning on the set of satisfied and neglected responsibilities allows a better understanding of what went wrong in the execution, and possibly pinpoints the root causes. This information could be used to improve the process, for instance, by regimenting some parts of the process or to make it more permissive. A responsibility relation is formally denoted as 𝑅(𝑥,𝑢,𝑣) where 𝑥 is a role, 𝑢 is a context condition, and 𝑣 is the task assigned to 𝑥. Intuitively, 𝑅(𝑥,𝑢,𝑣) states that any actor playing the role 𝑥 will be receptive to the request/need of bringing about 𝑣 if 𝑢 holds. Condition 𝑢 and task 𝑣 can both be simple activities or temporal patterns of activity executions. Example (Alignments and Responsibilities). In this example, we consider a process for coworking space reservation. When a reservation is made (Make Reservation – MR), then it can either be paid directly (Payment – P), or a request for remote payment is first sent (Send Payment Request – SPR), then a reservation confirmation upon payment is sent (Send Reservation Notification - SRN). Following this specification, two model runs are possible: ● 𝐸1 = <𝑀𝑅,𝑃 > ● 𝐸2 = <𝑀𝑅,𝑆𝑃𝑅,𝑆𝑅𝑁 > 25 Let us consider the observed execution trace: 𝑇 = <𝑀𝑅 > that is, only Make Reservation is observed. The possible alignments with trace 𝑇 are the following, where ≫ represents a mismatch (i.e., a move where either the log trace or the model moves one step). 𝐴1 =𝑀𝑅 ≫ 𝑀𝑅 𝑃 𝐴2 = 𝑀𝑅 ≫ ≫ 𝑀𝑅 𝑆𝑃𝑅 𝑆𝑅𝑁 Classical approaches conclude that A1 is the optimal alignment, having one mismatch only, while A2 has two. Therefore, the model execution closer to trace T is E1. Let us now assume that the model is complemented with an explicit representation of responsibilities, and that the employee is responsible for sending a Reservation notification only after the Payment Request has been sent and only in case it is sent. Assessing the two alignments against such a responsibility allows us to observe that the lack of Send Reservation Notification (SRN) in A2 is justified by the fact that the Payment Request was not sent (for some reason). Therefore, while the absence of Send Payment Request (SPR) is not justified by any responsibility (and thus has to be counted as a mismatch in the alignment), the absence of Send Reservation Notification (SRN) is justified by the non-occurrence of Send Payment Request. As a result, the two alignments can be considered equivalent in terms of the number of mismatches. In other words, the two process executions are equally possible as an alignment of the execution 𝑇. To capture these kinds of conditions in responsibilities, it is possible to rely on several temporal logics. In our proposal, we rely on Precedence Logic (Singh and Huns, 2005) for its simplicity. Precedence logic is an event-based linear temporal logic, obtained from propositional logic augmented with the temporal operator (⋅) before. Example (Responsibility). The responsibility mentioned in the previous example can be expressed by 𝑅(𝐸,𝑆𝑅𝑁,𝑆𝑃𝑅⋅𝑆𝑅𝑁) modeling that the employee 𝐸 is responsible that whenever Send Reservation Notification is performed, it is performed only after Send Payment Request (𝑆𝑃𝑅⋅𝑆𝑅𝑁). For details on the formalization, please refer to (Baldoni et al., 2024). 3.3 Process Alignment Process alignment corresponds to comparing a process path against an execution trace (i.e., an actual execution). Generally, the objective is to find, among the possible ones, an alignment that is optimal to a criterion of preference. Intuitively, an alignment proceeds step-by-step on the model and on the execution: at each step, if the activity in the model and the one in the execution match each other, a synchronous move is made, and both model and log advance one step. Otherwise, either the model moves and the log does not, or the other way around, the log moves and the model does not. Usually, to find an optimal matching, a cost function associated with mismatches (i.e., asynchronous moves) is defined. So, an optimal alignment is the one that minimizes the cumulative cost of the mismatches. Among the existing approaches, the classical one is to assign the same cost to each mismatch, so optimality is reached by minimizing the number of asynchronous moves. In our approach, an optimal alignment is determined by taking into account both the alignment between an execution and a model path and the involved responsibilities. Intuitively, we collect all the responsibilities that are involved in the process, and verify whether they are satisfied during 32 ideal flow, the company could identify where employees were deviating from the prescribed process and why. 4.4 Process Mining Application: a set of Guidelines Several companies offering support for process mining application exists. However, the cost for such services is accessible for big companies only. In this section we provide a few guidelines aiming at assisting small and medium-sized companies and organization in taking advantage of process mining. However, we are aware that the task remains challenging. Step 1 - Objectives and Data ● Define the objective of the analysis (business goal). ● Understand what data is available, what it represents, and how it can be integrated if multiple sources are used. Step 2 - Process ● Identify the process to analyse (better to focus one process at a time) Step 3 - Interest ● Identify a set of questions the company wants to answer ● Prioritize the questions in an ordered list Step 4 - Data preparation ● Select and prepare the data ● Make sure the data is consistent and assess the level of completeness ● If possible, fill the gaps by integrating with additional data sources Step 5 - Tools ● Identify which tools are suitable for answering the identified questions ● Apply the tools to the data ● Check which questions can be answered and which not Step 6 - Include the experts ● Analyse the results with the experts as soon as possible Step 7 - Iterate 33 ● Concrete answer to the questions may indicate that the questions were not well formulated or were imprecise. New questions may be triggered by the analysis, as well as interpretations of the results. ● Formulate new questions or refine the existing and repeat the analysis process 4.5 Strengths and Limits of the Demonstrator The approach and the methodology are very powerful and proved effective in a real-world scenario. The limit, in this case, stands in the lack of a process-oriented mentality in companies. Sometimes such a lack is fostered by the adoption of tools that, for meeting the needs of many realities, provide loose constraints on the contexts when things should be done. People adapt and invent (sometimes individual) ways/practices for adding the meanings that the tool does not incorporate. The education to a process-oriented mind is the actual effort. 5. Conclusions In this deliverable we reported on a number of tools that can serve the goal of contrasting depopulation in the mountain regions, and that can easily be applied to other geographic areas of interest. Our claim is that depopulation can be contrasted by fostering social innovation, and in particular by supporting the creation of smart collaborations in a broad sense, that is, both within a company and across companies with the help of technology. This is a complex goal, that requires the interplay of economic, social, and environmental actions. We have focussed on the impact of networks of co-working spaces, which can serve a wide range of purposes, and displayed a showcase of demonstrators that can support decision makers, politicians, and also companies in deciding about possible investments because they allow estimating the impact. 34 Bibliography M. Baldoni, C. Baroglio, E. Marengo, R. Micalizio (2024). Reasoning on responsibilities for optimal process alignment computation. Data Knowl. Eng. 154: 102353. J. Burgers-Pas, C. Seckler (2022). How to Make Smart Collaboration Work in Multidisciplinary Teams. In: Drechsler, A., Gerber, A., Hevner, A. (eds) The Transdisciplinary Reach of Design Science Research. DESRIST 2022. Lecture Notes in Computer Science, vol 13229. Springer, Cham. https://doi.org/10.1007/978-3-031-06516-3_20. R. Bouncken, M. Ratzmann, R. Barwinski, and S. Kraus (2020). Coworking spaces: Empowerment for entrepreneurship and innovation in the digital and sharing economy. Journal of Business Research 114, 102–110. M. Dumas, M. La Rosa, J. Mendling, H. A. Reijers (2018). Fundamentals of Business Process Management, Second Edition. Springer, ISBN 978-3-662-56508-7, pp. 1-527 H. K. Gardner (2016). Smart collaboration: how professionals and their firms succeed by breaking down silos, Boston, Massachusetts: Harvard Business Review Press. E. J Malecki (2011). Connecting local entrepreneurial ecosystems to global innovation networks: open innovation, double networks and knowledge integration. International Journal of Entrepreneurship and Innovation Management 14, 1, 36–59. Fluxicon Disco. Discover your process. https://fluxicon.com/disco/ Repast-simphony https://repast.github.io/repast_simphony.html. Repast Simphony Reference Manual. Available at https://repast.github.io/docs/RepastReference/ Repast Reference.html. A. Saraceni (2025). Sistema multiagente georeferenziato per la simulazione degli Spazi di Coworking, Andrea, tesi di laurea in informatica, Università degli Studi di Torino. M. D. Marino, A. Rehunen, M. Tiitu, and K. Lapintie (2023).New working spaces in the Helsinki Metropolitan Area: understanding location factors and implications for planning. European Planning Studies, vol. 31, no. 3, pp. 508–527. D. Grazian (2020). Thank God it’s Monday: Manhattan coworking spaces in the new economy. Theory and Society, vol. 49, no. 5, pp. 991–1019. I. Mariotti, M. Akhavan, and F. Rossi (2023). The preferred location of coworking spaces in Italy: an empirical investigation in urban and peripheral areas. European Planning Studies, vol. 31, no. 3, pp. 467–489. R. Belk (2014). You are what you can access: Sharing and collaborative consumption online. Journal of business research, vol. 67, no. 8, pp. 1595–1600. 35 E. Risi and R. Pronzato (2021). Smart working is not so smart: Always-on lives and the dark side of platformisation. Work Organisation, Labour & Globalisation, vol. 15, no. 1, pp. 107–125. J. Howaldt, M. Schwarz (2010). Social Innovation: Concepts, research fields and international trends. IMO International Monitoring. M. Grecchi, A. Colucci, L. E. Malighetti, F. Speciale (2025). Reasons and Challenges for the Regeneration of Abandoned Mountain Villages. In: Reinventing Mountain and Rural Villages. SpringerBriefs in Applied Sciences and Technology. Springer, Cham. https://doi.org/10.1007/9783-031-84764-6_1 M. P. Singh, H. N. Michael (2025). Service-oriented computing - semantics, processes, agents. Wiley, ISBN 978-0-470-09148-7, pp. I-XXXVIII, 1-549 M, Alderighi C. Baroglio, M. Chiesa, T. Ciano, C. Feder, V. Figini, E. Marengo, S. Tedeschi, Towards a Network of Co-working Spaces for Social Innovation in Mountain Areas, 32nd International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises, WETICE 2024, Reggio Emilia, Italy, June 26-28, 2024, p. 44-49, IEEE, 2024. Deliverable 5 - Best practices: Guidelines on: organisational change; firm performance; workers’ wellbeing; tools and devices; and digital compliance. URL: https://zenodo.org/records/14012280 [Autori: UniVda; UniTo; PoliTo; UniBas; LINKS]