This paper presents a heuristic structural optimization method using a genetic algorithm (GA) for the design of layered sound-absorbing structures with predefined frequency characteristics. The design model employs a symbolic knowledge representation in the form of a genome that encodes the number of layers, their materials, and thicknesses. This encoding enables the use of evolutionary mechanisms such as inheritance, mutation, and selection, typical of genetic algorithms. The evaluation of solution fitness was carried out through a semi-automated process integrating custom evolutionary software with the Norflag acoustic simulation tool. The proposed approach addresses an inverse design problem under constrained computational resources and limited simulation automation. The algorithm's effectiveness was confirmed experimentally—near-optimal structures matching the reference pattern were obtained within just 28 generations. This work contributes to the application of artificial intelligence in engineering design by demonstrating a logic-based, algorithmic reasoning process and highlighting the potential of heuristic methods in solving complex, multi-parameter optimization tasks.