Modern OFDM communication systems must have the property of adapting to the changing parameters of the communication channel. Intelligent decision support systems using UFMS technology can effectively solve this problem. The use of UFMS filtering allows you to select the optimal value of the cyclic prefix and increase the spectral efficiency of the OFDM signal. However, this increases the computational cost of performing digital filtering procedures, which leads to an increase in network latency. Therefore, the development of a fast algorithm for performing numerical-theoretical signal transformations for OFDM systems using UFMC technology is an urgent task.