On a high-dimensional objective genetic algorithm and its nonlinear dynamic properties

The revival of multi-objective optimization is mainly resulted from the recent development of multi-objective evolutionary optimization that allows the generation of the overall Pareto front. This paper presents an algorithm called HOGA (High-dimensional Objective Genetic Algorithm) for high-dimensional objective optimization on the basis of evolutionary computing. It adopts the principle of Shannon entropy to calculate the weight for each object since the well-known multi-objective evolutionary algorithms work poorly on the high-dimensional optimization problem. To further discuss the nonlinear dynamic property of HOGA, a martingale analysis approach is then employed; some mathematical derivations of the convergent theorems are obtained. The obtained results indicate that this new algorithm is indeed capable of achieving convergence and the suggested martingale analysis approach provides a new methodology for nonlinear dynamic analysis of evolutionary algorithms.

Research highlights
► In this paper, we propose an algorithm called HOGA (High-dimensional Objective Genetic Algorithm) for high-dimensional objective optimization on the basis of evolutionary computing.
► We employ the martingale analysis approach to analyze the nonlinear dynamic property of HOGA.
► Our work provides a new way for nonlinear analysis of evolutionary algorithms.