Article Preview
TopIntroduction
In 5G and beyond (B5G) systems, the emerging new service scenarios and business models driven by user demands will differ from today’s in terms of deepened connections and extended service chains (Yu et al., 2020; Chien et al., 2019; Hou et al., 2017). Traditional generic service models are no longer appropriate for highly customized and intelligent services (Aazam et al., 2019; Van Hee et al., 2001). Diversity and dynamics are the new features of user demands, which require the service providers to reinvent their service models in accordance with fast-changing demands. For example, Figure 1 shows a change occurring in the service process in the tourism industries during and after the coronavirus disease 2019 (COVID-19) era. To protect public health, a new regulation requests that tourists send both health code and journey data within 14 days to the travel agencies (Ienca et al., 2020). To obtain the health information of each applicant, a service designer needs to add two new operations “Check Health Code” and “Check Journey Data” after the operation “Receive Inquiry”; see the dotted box in Figure 1. Reinventing service models involves allocating resources to the newly added tasks to rapidly adapt to market changes (Van Hee et al., 2001; Mendling et al., 2018). Additionally, a sudden burst of user requests requires the service providers to add resources to accommodate the increased load conditions, e.g., the newly initiated service requests for insurance claim during or after the occurrence of disasters (Doan et al., 2019; van der Aalst et al., 2007). This also involves the problem of allocating newly added resources to pending tasks to respond to user service requests.
Figure 1. An example of service process changes in the tourism industries during and after COVID-19 era, where the part highlighted in red is a new addition in response to COVID-19
It is not easy to develop a simple and effective approach to allocating resources globally optimally to quickly satisfy new business service requirements in future B5G era. The reason is that adding resources to task nodes with no growth in others does not improve performance, e.g., the number of users that can be served per unit time and the average response time per request. Generally, a service consists of one or more business processes, whose activities work together to deliver the specific service according to internal business rules. A business process model defines the activities to be executed, their data objects and resources, and the execution order of activities (Combi et al., 2009; Natschläger et al., 2015). Each activity has the important attribute, i.e., duration, which specifies the allowed temporal spans of the activity (Combi et al., 2009), and requires a certain number of resources to complete. The appropriate allocation of resources to an activity has a direct impact on the performance of entire service processes, e.g., the cycle time of process instances (Sheng et al., 2009) and the data quality of information systems (Liu et al., 2020). However, the optimal resource allocation under constraints, as a typical problem of operations research (OR), is NP-hard and difficult to solve (Doan et al., 2019; Xu et al., 2019; Ma et al., 2020).