Collective Adaptation in Smart Cities

Antonio Bucchiarone, Annapaola Marconi, Naranker Dulay, Anna Lavygina and Alessandra Russo

The Internet of Services (IoS) foresees a future Internet in which the provisioning of, access to and use of services will be as widespread as content is today. Smart Cities are becoming one of the main drivers in the eruption of this wave. The urgent need for a more efficient and sustainable society, together with the spread of ubiquitous communication networks, highly distributed wireless sensor technology, and intelligent management systems, makes the Smart City ecosystem an ideal ground for IoS.
In this setting, the role of service-oriented computing is to enable the integration and interplay between new and legacy city services to solve current and future challenges and support the creation and delivery of innovative and efficient services for the citizens.

Smart Children Mobility

Figure 1: A partial overview of the smart children mobility system.

A key challenge that still needs to be overcome for this to become a reality, is the capability of dealing with the continuously changing and complex environment in which Smart City applications operate. Consider for instance the case of a smart children’s mobility system (depicted in Figure 1), supporting service users (parents, children) and providers (drivers, teachers, traffic aids, volunteers) in their daily operation and management of children mobility services (e.g., school buses, walking buses, bike trains, ride-sharing among parents).
If the aim is to deliver smart children mobility services to citizens, all the entities involved cannot be operated each by itself, but should become part of an integrated mobility solution, the Smart Children Mobility System (SCMS), that supports users (parents, children) and providers (teachers, drivers, traffic aids and volunteers) in their daily operation and management of the different mobility services.

Even though these entities are generally autonomous, they dynamically form collaborative groups, called ensembles, to gain benefits that otherwise would not be possible. The example of such an ensemble is a Walking Bus Route (WBR) (see Figure 1) which coordinates the adaptation behavior of multiple entities (WSB Manager, Route Manager and Volunteer Management) and in return gives them certain benefits (e.g., safe and dynamic handling of children walking bus routes). Membership of an ensemble may temporarily reduce the flexibility of its entities. Within this context, isolated entity self-adaptation is not effective. We can easily image what happens if a volunteer assigned to a specific WBR and then silently changes her mind and decides not to travel. It is likely to cause the route cancellation if notified in delay. Even more serious consequences arise if the weather conditions deteriorate and the WSB manager decides to suspend the Walking School Bus. In such a systems new approach for adaptation are therefore needed that allow (i) multiple entities to collectively adapt with (ii) negotiations to decide which collective changes are best. Collective adaptation also raises a second important challenge: which parts of the ensemble should be engaged in an adaptation? This is not trivial at all, since solution for the same problem may be generated at different levels. For instance, the volunteer cancellation can be resolved in the scope of the WBR (same ensemble), by finding a substitute, or in a wider scope finding an alternative way of bringing children to school (i.e., assign them to another walking bus, use school bus or RideSharing initiatives, etc..). The challenge here is to understand these levels and create mechanisms to decide the right scope for an adaptation for a given problem.

In ALLOW Ensembles project, we are realizing a novel approach for collective adaptation that is driven by awareness of the capabilities, goals, constraints and preferences of humans and entities, as well as the knowledge of the environment. Our adaptation process is distributed and is controlled by a multi-criteria decision making function that is combined with an analytic hierarchic process (AHP) to select best adaptation alternatives.