Genetic algorithm search space splicing particle swarm optimization as general-purpose optimizer

• Search space modification is presented to aid global convergence of PSO algorithms.
• GA and PSO was used for heuristic search space splicing and subspace exploration.
• GA-SSS-PSO could optimize all tested functions to an error approaching zero.
• This method was demonstrated to robustly handle the investigated chemometric data.

A heuristic search space splicing scheme has been implemented to aid the convergence of the particle swarm optimization (PSO) algorithm to the global optimum. Genetic algorithm (GA) was used to splice the search space into smaller subspaces, thereby reducing the number of local minima. PSO algorithm was subsequently used to locate the global optima in the subspaces. A set of 11 well-known test functions had been used for the assessment of this novel GA search space splicing PSO (GA-SSS-PSO) architecture. Of the methods tested in this study, the GA-SSS-PSO approach was the only one that could optimize all functions to a desirable level. To demonstrate the algorithm's applicability, three optimization tasks of different categories commonly faced in the field of chemometrics were subjected to optimization by GA-SSS-PSO and results indicated that the novel hybrid algorithm provided robust performance for both theoretical and real life problems and may be suited as general-purpose optimizer for medium-sized optimization tasks.