Special Issue on Self-integrating Systems: Mastering Continuous Change
• 大类 : 工程技术 - 2区
• 小类 : 计算机：理论方法 - 2区
The goal of this special issue if to highlight the current state-of-the-art in self-improving system integration and to foster novel developments. This corresponds to two closely coupled concepts: self-integration and self-improvement.
Self-integration: In engineering, the notion of integration describes a process in which several component (sub-)systems are brought together and interconnected into a unified system. This aims to achieve a correctly working unit, where the subsystems work together to provide desired functions, with acceptable performance and dependability properties. In classic engineering, the integration of (sub-)systems has been done at design-time, with rigid specifications and testing of (sub-)system interfaces and performance. However, with increasingly networked and open systems (e.g., internet of things, smart homes, cities and electric grids), we now face the challenge of integrating systems dynamically, as rapidly as possible. Because of the dynamic contexts - where goals, resources, and knowledge required for integration change rapidly - increasing efforts have been directed towards new processes and computations that allow intelligent systems to do most of the integration themselves. Hence, self-integration is defined as an ongoing autonomous process for linking a potentially large set of heterogeneous computing systems, devices, and software applications; so as to meet system goals. The linking itself is done physically or functionally. We consider this process to be continuous, i.e. it is never finished, as the way in which the integrated elements - software and hardware - act together must adapt to external changes, goal evolutions and autonomous decisions of these elements.
Within this process, two major tasks are performed: (i) connections between subsystems are established, tested, and assessed; and, (ii) the particular subsystems in combination with their behavioural strategies are configured depending on the operational and functional area. This has a strong overlap with the concept of "self-organisation": (i) connecting subsystems or elements refers to building the system's structure; and, (ii) configuration is mainly concerned with defining the desired behaviour. It also relies on concepts from self-aware computing systems that use internal knowledge and learning processes to reason about their resources and state. Multi-agent systems also provide relevant concepts in terms of organisational paradigms, where agents can cooperate or compete, forming hierarchies, coalitions, teams, federations, societies and so on.
Self-improvement: implies that self-integrating systems do not settle for the status quo, but rather rely on continuous learning to optimise their behaviours with respect to their goals; even though they may or may not be able to reach optimal states or behaviours.
In this special issue, we aim to summarise the current developments in autonomous and self-improving system-integration. We expect contributions to cover at least one of the following aspects of self-integration (although this list is by no means exhaustive, leaving the call open to other related contributions):
- Goal specifications and conflict management for self-integrating systems
- Performance and dependability aspects and quantification methods
- Learning techniques for self-improving self-integration
- Learning techniques that work with sparse feedback and/or prior knowledge
- Modelling expected behaviour and mutual influences in system-of-systems constellations
- Real-time prioritization; representation of relevancy and importance
- Runtime model integration (i.e., Models@runtime) also takling timing into consideration (i.e., real-time restrictions)
- Decision and planning techniques for self-adapting the integration status at runtime
- Mechanisms for guiding the behaviour of autonomous entities without direct intervention (e.g. using norms)
- Runtime system verification and validation
- Security issues and guarantees in self-integrating systems
- Testbeds and benchmarks for self-integrating systems