Peak factor optimization of multi-harmonic signals using artificial bee colony algorithm

Multi-harmonic signals are useful in many applications to estimate the frequency response of a device. To minimize noise effects, signal power should be maximized within the amplitude limitations of the signal generation and acquisition hardware. Signal power depends only on the harmonic amplitudes, while the signal amplitude also depends on the harmonic phases. Careful choice of these phases allows the minimization of the peak factor which corresponds to the minimization of the signal's amplitude for a given signal power. Multiple methods have been proposed to minimize a signal's peak factor, including closed form expressions and iterative algorithms. However, the existence of many local extrema suggests the use of global search algorithms. This paper proposes the use of the artificial bee colony algorithm. The solution is further improved by using a Nelder-Mead Simplex algorithm. The artificial bee colony algorithm results are compared with those of previously published methods.