Artificial Plants With A Dynamic Population Algorithm

Artificial plant optimization algorithm (APOA) is a novel population-basedstochastic algorithm. Inspired by photosynthesis phenomenon, phototropismphenomenon and apical dominance phenomenon, APOA designs photosynthesisoperator, phototropism operator and apical dominance operator. In this article,the following research work is provided:In nature, new bifurcations are produced in each branch to increase theefficiency of photosynthesis phenomenon. Furthermore, due to the weather andother unpredicted factors, some branches will be dead. Inspired by both of them,a new variant, artificial plant optimization algorithm with dynamic population(APOA-DP) is designed in which some new branches will be rooted while somebranches will also be dead. To test the performance, several benchmarks areselected, APOA-DP achieves the better performance when compared with thestandard version.In APOA-DP, the dead branches are selected according to their bad fitnessvalues, and the population diversity is decreased. With this manner, APOA-DPis easily trapped into the local optima. To avoid this problem, a new clustermethod is employed to enhance the performance of APOA-DP, and is called asartificial plant optimization algorithm with dynamic population and clustermethod (AOPA-DC). Furthermore, some parameters are tested by uniformdesign. Simulation results show APOA-DC is better than APOA-DP especiallyfor high-dimensional mutli-modal problems.Finally, artificial plant optimization algorithm with dynamic populationand cluster method is applied to solve nonlinear equation problems. Simulationsresults show it is effective.