Noise immunity enhancement by using a genetic-like coding algorithm for metric source without introduction of additional redundancy

The problem of increasing noise immunity for sending messages usually solved by introducing the additional redundancy or by re-transmission of distorted fragment of the message. At the same time, the resulting delay is not always acceptable. An original concept for irredundant code efficiency rising is discussed. It is shown that some genetic-like algorithm applying to the class of permutative equivalent codes can suggest some code that will be too much stable in respect of disturbance appearance. Source space topology, concordant metrization procedure, miscellaneous error structure can be properly taking into account in technical feasibility of this algorithm. Examples of 2D Euclidean source spaces and their coding are presented. Reduction achieved for average displacement of decoding is at least 32% better than that for average values or at least five standard deviations.