A dynamic programming algorithm for constructing optimal "1"-ended binary prefix-free codes

The generic Huffman-Encoding Problem of finding a minimum cost prefix-free code is almost completely understood. There still exist many variants of th is problem which are not as well understood, though. One such variant, requ iring that each of the codewords ends with a "1," has recently been introdu ced in the literature with the best algorithms known for finding such codes running in exponential time. In this correspondence we develop a simple O( n(3)) time algorithm for solving the problem.