Description Approach And Decoding Probability Computation Approach For Random Network Coding

The most clearly differences between Multi-User Information Theory and Network Information Theory is that, besides Source Channel, the research of intermediate nodes is also the scope of Network Information Theory. Research shows that throughput of a network can be increased by applying network coding. As a new branch of Network Information Theory, network coding will be a breakthrough of the development of communication technique.Random network coding, an important approach to construct network coding, satisfies the need in the practice application. This thesis first study the decoding probability under random network coding with a model of wireless multicast network. The advantage of random network coding is that the network topology is not necessary to intermediate nodes. But it brings a new problem, that is the sink node may not ensure the decoding is entirely correct. This thesis gives a method to calculate the exact decoding probability of sink node in a multicast network model. Then this thesis gives a method to describe random network coding in an arbitrary wireless network. The method is based on differential equations and the solution of the differential equations is the innovative packets each node possess over time, and the results show that the simulated curves match the analytical solutions well.