An agent-based clustering framework for reliable satellite networks

Agent technique is applied to deal with varying demands and satellites are specifically modeled as agents with autonomy and responsiveness. While most studies introduce efficient agent-based satellite networks, the efficiency of satellite operation that reflects network performance needs to be quantified. Moreover, it is still difficult to configure and manage large-scale networks with existing frameworks as a result of resource limitations. This paper proposes a practical framework for satellite networks to solve these issues by means of agent-based modeling, clustering technique and game theory. An agent hierarchical structure including three control levels is designed for single satellite to manage components. According to the structure, a novel indicator named effectiveness is presented for efficiency assessment. A clustering network framework is proposed to guarantee the stability and scalability, leading to lifetime maximization in resource–critical situations. Furthermore, a collaborative routing scheme based on the Nash bargaining solution is introduced to obtain the globally optimal route. Case analysis shows that the superiority of the proposed framework compared with state-of-the-art algorithms in terms of reliability improvement and energy efficiency while maintaining low network delay.