Optimised constructions for variable-length error correcting codes

An optimised construction of variable-length error-correcting (VLEC) codes is proposed. Compared to the state-of-the-art, the three main improvements consist in: (1) improving the codewords search algorithm complexity/efficiency trade-off; (2) loosening the codewords deletion rule; (3) taking advantage of eventual previous searches. It is shown that the optimised algorithm can find good VLEC codes for alphabets of up to 200 symbols, and outperforms existing algorithms.