Actually, a center-fed dipole that is slightly shorter than 1/2 wavelength (using the speed of light in a vacuum) might be considered "optimal". The reason why this is so, is that for such a dipole, the capacitative and inductive reactances cancel each other, making the impedance of the antenna purely resistive.
When the antenna has purely resistive impedance, the apparent power supplied to the antenna is equal to the true power supplied. An antenna for which this is not the case can also be an efficient radiator if the inductive or capacitative reactance is canceled by an external capacitor or inductor. Such an arrangement may be slightly less than ideal, however, because there will be circulating current between the antenna and the component used to balance the impedance. That is, there will be energy that is passed back and forth between the antenna and the component used to balance the impedance. Each time that energy is passed back and forth between the antenna and the balancing component, some energy is lost. This lost energy, in some cases, could be a significant fraction of the energy one is hoping to receive from the antenna, or hoping to transmit via the antenna. That, ultimately, is why an antenna which "naturally" has no reactive impedance at the receiving/transmitting frequency might be considered "optimal".
The following graph, taken from Wikipedia shows the reactive impedance going crossing 0, at a point slightly below 0.5 wavelength. Notice that for antennas significantly shorter than that, the resistive impedance gets very small, while the (absolute) magnitude of the reactive impedance gets very large.
