An algorithm for generating n Huffman codes with a worst case of (n log/sub 2/ n+n log/sub 2/ log/sub 2/ n) comparisons

The method is based on a Heapsort algorithm for sorting a sequence of n data elements with a worst case of (n log/sub 2/ n+n log/sub 2/ log/sub 2/ n) comparisons. The idea of heap restoring in the Heapsort algorithm is incorporated into the Huffman tree building process. The performance of the proposed algorithm in execution time is evaluated with a list of ordered, random, and reverse ordered data containing from 5000 to 500000 items.