Rainbow Network Flow of Multiple Description Codes

This paper is an enquiry into the interaction between multiple description coding (MDC) and network routing. We are mainly concerned with rate-distortion optimized network flow of an MD source from multiple servers to multiple sinks. Maximizing a collective metric of the MD source reconstructed at all sinks, constrained by edge capacities, is a very different problem from conventional maximum network flow. The objective function involves not only the flow volume but also the diversity of the flow contents (distinct MDC descriptions), hence the term of rainbow network flow (RNF). The RNF problem is also closely related to lossy network coding. For general network topology, general fidelity function, and an arbitrary distribution of MDC descriptions on the servers, we prove that the RNF problem is Max-SNP-hard, i.e., there is no polynomial-time algorithm to even approximate the optimal solution with arbitrarily good approximation ratio, unless P=NP. However, the problem becomes tractabl...

• Type: Journal
• Year: 2008
• Where: TIT
• Authors: Xiaolin Wu, Bin Ma, Nima Sarshar
• Keywords: Descriptions, MDC Descriptions, Network Flow, TIT 2008