AbstractGenetic algorithms (GAs) have been recently applied to uncoded space‐time labeling diversity systems (USTLD) to improve upon diversity in wireless links. GAs were used to optimize the secondary mappers that are used to encode the information to be transmitted. Existing literature only achieved a local optimal design, at the fraction of the computational costs. Hence, this article focuses on improving upon the GAs for USTLD systems by merging it with a global‐local search technique, called the global neighborhood algorithm (GNA). The GNA algorithm was tested on M‐QAM, M‐PSK, M‐APSK, and asymmetric M‐APSK constellations of modulation order M , and . In the case of M‐QAM constellations, the GNA was able to match, but not improve upon existing mapper designs. In the case of the 64PSK constellation, the GNA was able to produce mapper designs that improved upon the error performance of existing mapper designs by approximately 1 dB. In the case of the 32APSK constellation, the GNA was able to improve upon the error performance of existing mapper designs by approximately 0.5 dB. Moreover, the GNA produced mapper designs for the 11+5APSK and asymmetric 16APSK constellations that improved upon the error performance of existing mapper designs by approximately 0.5 and 4 dB respectively. Finally, the computational complexity of the proposed GNA algorithm was analyzed and the algorithm was shown to be which is slightly more complex than the initial GA approach with complexity , but significantly less complex than the enhanced GA with complexity . The increased complexity also came with significant gains such as improved LD mapper designs.