Conventionally optimal control theory has been used in the theoretical design of laser pulses through the direct variation in the electric field of the laser pulse as a function of time. This often leads to designed laser pulses which contain a broad and seemingly arbitrary frequency structure that varies in time in a manner which may be difficult to realize experimentally. In contrast, the experimental design of laser pulses has used a genetic algorithm (GA) approach, varying only those laser parameters actually available to the experimentalist. We investigate in this paper the possibility of using GA optimization methods in the theoretical design of laser pulses to bring about quantum state transitions in molecules. This allows us to select only a small limited number of parameters to vary and to choose these parameters so that they correspond to those available to the experimentalist. In the paper we apply our methods to the vibrational-rotational excitation of the HF molecule. We choose a small limited number of frequencies and vary only the associated electric field amplitudes and pulse envelopes. We show that laser pulses designed in this way can lead to very high transition probabilities.