Goal-oriented Holonics for Complex System (Self-)Integration: Concepts and Case Studies

System integration from sub-systems has always been a major engineering problem, which is progressively exacerbated by (1) sub-systems becoming more diverse, self-* and autonomous (2) systems operating in open environments, with third-party sub-systems joining and leaving unpredictably, (3) system (self-)integration being an ongoing process, increasingly needed at runtime. The fact that this problem occurs more and more often, as systems are built increasingly by composing existing sub-systems, requires rigorous, reusable integration solutions to replace ad-hoc approaches. In a complex world of uncertainty and change the new system integration paradigm must feature two main characteristics: support for a system-of-systems approach to manage complexity, and support for a high-level relation between sub-systems to manage diversity, uncertainty and dynamics. We propose a conceptual modelling solution combining holonic principles with goal-based relations. We highlight the key properties of holonic designs that support a systems-of-systems approach. We then specify the high-level relations between holonic sub-systems as goal-oriented requests and replies. Argumentation is grounded via concrete examples from existing complex systems. The proposed paradigm paves the way for future methodologies and tools for designing the next generation of socio-technical and cyber-physical systems.