Providing Internet service above the clouds is of ever-increasing interest and in this context aeronautical {\it{ad-hoc}} networking (AANET) constitutes a promising solution. However, the optimization of packet routing in large ad hoc networks is quite challenging. In this paper, we develop a discrete $\epsilon$ multi-objective genetic algorithm ($\epsilon$-DMOGA) for jointly optimizing the end-to-end latency, the end-to-end spectral efficiency (SE), and the path expiration time (PET) that specifies how long the routing path can be relied on without re-optimizing the path. More specifically, a distance-based adaptive coding and modulation (ACM) scheme specifically designed for aeronautical communications is exploited for quantifying each link's achievable SE. Furthermore, the queueing delay at each node is also incorporated into the multiple-objective optimization metric. Our $\epsilon$-DMOGA assisted multiple-objective routing optimization is validated by real historical flight data collected over the Australian airspace on two selected representative dates.
翻译:在云层之上提供互联网服务越来越令人感兴趣,在此情况下,航空上的 liit{ad-hoc} 网络(AANET) 是一个很有希望的解决办法。然而,大型特设网络中包路程的最佳化是相当具有挑战性的。在本文中,我们开发了一个离散的美元多目标遗传算法($\epsilon$-DOMGA),以共同优化端到端的悬浮、端到端的光谱效率(SE)和路径到期时间(PET),其中具体说明了路线路线在不重新优化路径的情况下可以依赖多久。更具体地说,专门为航空通信设计的远程适应编码和调节(ACM)计划被利用来量化每个链接的可实现的SE。此外,每个节点的排队延迟也被纳入了多目标优化度指标中。我们美元-DOMGA 协助的多点路程优化得到了两个选定代表日期在澳大利亚领空上收集到的真实历史飞行数据的验证。