A case study of innovative population-based algorithms in 3D modeling: Artificial bee colony, biogeography-based optimization, harmony search

Deterministic or analytical methods for computing the global optima of a functional have been extensively applied in a wide range of engineering applications. Nevertheless, it is wellknown they usually lack of effectiveness when dealing with complex nonlinear optimization problems. In particular, such a shortcomings have been addressed by using approximate approaches, named metaheuristics. Among them all, those methods using a population-based scheme, e.g. the evolutionary algorithms, have been the most successful optimization strategies. Recently, innovative population-based algorithms such as ABC, BBO, and HS have arisen as promising optimization methods due to they provide a good tradeoff between design and performance when compared to other more elaborated methods. In this work, we aim to first introduce the particular design of these three cutting edge algorithms, and additionally analyse their performance when tackling a challenging real-world optimization problem. In particular, our case study of numerical optimization tackles a computer vision problem named 3D range image registration for 3D modeling tasks. Computational experiments have been conducted comparing the performance of ABC, HS, and BBO against other contributions in the state-of-the-art of 3D image registration.