Presentation of the dissertation project and the research results to date (Long)

Presentation of the dissertation project and the research results to date:

Background and Introduction of the Research Project

This project currently contains mainly two parts. The heuristic algorithm based optimization of multimodal optimization of satellite networks and the Multi-objective QoS Quality of Service) optimization of multimodal routing information of satellite networks. The two parts are as follows:

1. Due to the recent developments in wireless technology and electronics, it is feasible to develop pervasive algorithms for satellite environments. Satellite networks consist of different satellites in different orbits connected by inter-satellite links (ISLs). They can provide service for all kinds of users located in the air, above the offing and on various areas on the earth’s surface with complex geographical structure. Divided by the orbit altitude, there are low earth orbit (LEO), medium earth orbit (MEO) and geostationary earth orbit (GEO) satellite networks. Compared with those LEO, MEO or GEO single-layered satellite networks, the LEO/MEO multi-layered satellite network has a low link congestion and high robustness and is becoming a critical issue in the research on satellite networks. Since LEO/MEO satellite networks are very different from traditional ground networks with rapidly changing topology and complex structures, the QoS routing of MLSN is always a difficult problem.

To solve the QoS routing of MLSN, multimodal routing information is adopted to satisfy the QoS requirements of the users and the dynamic topology of satellite networks. Though dozens of routing protocols dedicated to satellite networks have been proposed in the past decades, only a few of them have been actually launched in operating satellite networks to verify their quality of service. Even in simulation environments, each QoS requirement is considered or optimized separately. For example, some papers only consider the end to end delay, others only consider the link utilization. And in some papers, the throughput is the only focused requirement. In one word, all these previous routing protocols are single mode. These single-mode routing protocols have obvious disadvantages. When some QoS requirements are considered, others are omitted intentionally or accidentally. So this project imports a multi-mode routing protocol to improve the performance of the routing protocol. Heuristic algorithms are used to optimize the multimodal QoS requirements. An ant colony algorithm, a genetic algorithm, a beehive algorithm and a taboo search algorithm are included and experiments are done to test the performance of the new multimodal routing protocol.

2. Compared with the LEO, MEO or GEO single-layered satellite networks, a multi-layered satellite network combined of different satellite layers can yield a much better performance. Therefore, multi-layered satellite IP networks have recently become a research hotspot. One of the challenges in multi-layered satellite networks is the development of specialized routing schemes and algorithms. Since the resources in satellites are constrained, the routing scheme should be simple and efficient. It should simplify and reduce the computational load of the satellite. The routing algorithms should compute paths with a low communication overhead and satisfy all the users’ QoS requirements. At the same time, load balance and congestion avoidance are also required for a good satellite network routing algorithm. In order to make the routing protocol efficient, multimodal QoS requirements need to be optimized in a simple and efficient way. This project uses a proper equal constraint (PEC) method and order method to optimize the multi-mode routing information while make the routing protocol efficient.

Results and conclusions of the Research Project

In part one of the projects, the multimodal QoS routing scheme using heuristic algorithms is adopted in order to provide QoS guarantees for LEO/MEO MLSN. Before applying the heuristic routing algorithms, the virtual topology grouping strategy is first improved and less time slots in a system cycle are obtained which makes the cost of switching between different snap shots much lower and guarantees the performance of the heuristic routing algorithms. Compared with the SPF algorithm, the heuristic based multimode QoS routing algorithm can provide better service in link congestion, package loss and calling block especially when the load is high. However, the heuristic algorithm needs a period time to converge. How to shorten the convergence time of the heuristic algorithms will be an important issue for us to research in our future work. The relation between choosing parameters and the quality of the result are also involved in our future research.

In part two of the projects, we proposed a multimode QoS routing algorithm based on the PEC and Order optimization method. Compared with the previous SPF routing algorithm, it can satisfy all the QoS requirements of the users. It will balance the network traffic load and avoid package loss to some degree. The algorithms run on a terrestrially based station and will respond according to the arrival of the user applications. This way, the powerful computing and storage resources of the ground base station can be applied. This frees the satellite routing algorithm from the constraints of the resources on board the satellite. Shown in our simulation results, our algorithm performs better than SPF. The comparison of our algorithm and the heuristic algorithms will be included in our future work, for heuristic algorithms can also solve the problem of multi-QoS optimization routing.