Adaptive vs. self-adaptive parameters for evolving quantum circuits

Abstract

Setting the values of various parameters for an evolutionary algorithm is essential for its good performance. This paper discusses two optimization strategies that may be used on a conventional Genetic Algorithm to evolve quantum circuits: adaptive (parameters initial values are set before actually running the algorithm) or self-adaptive (parameters change at runtime). The differences between these approaches are investigated, with the focus being put on algorithm performance in terms of evolution time. When taking into consideration the runtime as main target, the performed experiments show that the adaptive behavior (tuning) is more effective for quantum circuit synthesis as opposed to self-adaptive (control). This research provides an answer to whether an evolutionary algorithm applied to quantum circuit synthesis may be more effective when automatic parameter adjustments are made during evolution. © 2010 Springer-Verlag Berlin Heidelberg.