Dynamic programming multi-resource allocation and its applications to vector quantization

This paper proposes a new multi-resource allocation algorithm for the design of vector quantizers and image coding systems. The algorithm iteratively performs the optimal allocation one resource at a time using a dynamic programming technique. The algorithm can utilize the resources efficiently and allows the input sources to the quantizers in the systems to have different distribution statistics. The algorithm is applied to construct various coding systems. Simulation results show that the algorithm can effectively allocate the resource available to the quantizers in the systems so that the performances of the systems are better than that of the systems designed by other allocation algorithms.