A Novel Optimization Algorithm: The Forest Algorithm

A new optimization algorithm, namely the Forest Algorithm (FA), is introduced for the first time. This algorithm simulates trees´ growth, reproduction and death in a forest to perform optimization. In the algorithm, trees and branches represent a collection of trial solutions and parameters needed to be optimized respectively, and three mechanisms, i.e. Growth, proliferation and death, are employed for improving trees´ vitality, which is a factor defined to evaluate the fitness of trial solutions. This algorithm in general execute a global optimization by operating on a group of trial solutions in parallel, but its growth mechanism, which adopts a parameter sweeping method, is a local optimization, so it combines the ability to find global optima of the global optimization and the fast convergence of the local optimization. Several numerical experiments are conducted, in which the performance of the FA in terms of the global optimization capability, accuracy and efficiency is evaluated and compared to that of some widely-used global optimization algorithms such as the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO). Results shown the FA is able to perform global optimization effectively and with high accuracy.