A Meta-Model for Representing System-of-Systems Ontologies

A System-of-Systems (SoS) is a large-scale complex system that integrates multiple constituent systems, which have managerial and operational independence. In order to achieve higher-level common goals of an SoS, it is important to systematically integrate independent constituent systems by thoroughly analyzing and designing the target SoS as a whole. But before conducting these engineering activities, a number of various SoS stakeholders and engineers should be able to understand their SoS. In order to provide a holistic view as a common knowledge base, this paper focuses on developing a conceptual meta-model that represents SoS ontologies. By investigating several documents for Mass Casualty Incident (MCI) response systems, we identified essential objects and required features for SoS descriptions. Based on the investigation, we generalized the objects into SoS entities, and we develop a meta-model, called M2SoS (Meta-model for System-of-Systems). To design our M2SoS, we borrowed organizational concepts from meta-models for multi-agent systems, and entities and relationships are redefined to specify SoS concepts in M2SoS. Finally, M2SoS is analyzed with respect to SoS characteristics, and we evaluate if M2SoS can represent high-level ontologies for two SoS case scenarios.

1 Introduction

A System-of-Systems (SoS) is a type of large-scale complex system that utilizes collaborative capabilities from the integration of component systems. An SoS consists of multiple Constituent Systems (CSs) as components, which have managerial and operational independence [1], and CSs might be designed and developed to achieve their own independent system-level goals. Also, CSs could be heterogeneous systems and geographically dispersed, while moving towards agreed SoS-level common goals [12]. This type of systems has become more widely required in many different domains, such as national defense, climate observation, and disaster response domains. Unlike system engineering of other types of systems, CSs of an SoS are highly autonomous and they have more independent system-level features, so the integration of multiple CSs can lead to the extremely high complexity. For this reason, there have been a lot of needs to handle SoS-specific engineering issues, and recently a number of analysis, design, development and maintenance techniques have been actively researched [2], [8], [10].