۱- Introduction

۲- Game Theory

۳- Research Method

۴- Results

۵- Synthesis of Best Practice

۶- Discussion

۷- Conclusions

References

Abstract

Systems-of-systems (SoS) are becoming increasingly common in more and more domains, spreading from the initial focus on government-controlled areas such as defense to open market industries. This implies that collaborative SoS are becoming more important, where the constituents need to be given incentives to join and remain within the SoS. Game theory has been proposed as a framework to model and analyze such SoS mechanisms. It aims at providing incentives to the independently operated and managed constituents. This paper presents a systematic literature review on the applications of game theory to SoS engineering, together with a synthesis aiming at capturing the best practices for doing such an analysis. The main conclusions are that game theory can be applied to SoS in a wide range of application areas, and deal with problems related to acquisition, design, and operations. In particular, the operational formation of SoS are well suited for this kind of analysis, and it often requires the use of simulation techniques. However, most results in the field lack a validation in practice.

Conclusions

This paper has provided a systematic literature review in the area of GT applied to SoSE, and this contributes a starting point for researchers wishing to delve deeper into the area. It has also synthesized a best practice process for applying GT to more complex and realistic SoS problems, and this contribution is more towards practitioners. The main conclusions are that GT can be fruitfully applied to SoS in a wide range of application areas, and deal with various problems related to acquisition, design, and operations. In particular, problems related to the operational formation and dissolution of SoS are well suited for this kind of analysis. Due to the nature of SoS, that involve a large number of heterogeneous constituent systems that have a wide range of possible actions at their disposal, analytic techniques for evaluation do not suffice, but instead simulations are necessary. In particular, agent-based simulations are frequently used in the analysis. Unfortunately, the reported results from applying GT are not validated in practice in the literature, and there is still some way to go to properly ascertain the value of this as a practical SoSE technique. As for future research, a few areas would benefit from attention. One is to lower the threshold of experimenting with modeling and simulation. This could require better tools, but it could also take the form of libraries for those who wish to retain the flexibility of ordinary programming languages. On the theoretical side, deeper investigations into mechanism design as a technique of synthesizing the rules of an SoS would be highly interesting, as would the application of meta-game theory to the engineering processes that result in the SoS.