Scaled artificial bee colony programming

Abstract

Problems of symbolic regression aim to develop a function, described in symbolic form, that fits a given target fitcases. Artificial bee colony programing algorithm (ABCP) is one of the most feasible automatic programming methods that was proposed to solve symbolic regression problems. ABCP concept is based on swarm bee optimization and proved to achieve satisfactory performance. This paper proposes a novel variant of ABCP algorithm referred to as scaled artificial bee colony programming (SABCP). The SABCP variant results from structural modifications of ABCP which apply to food sources and solution representation as well as prediction model output form. The designed variant is checked for performance on symbolic regression problems and compared to other programming methods.