Double Shrinkage Algorithm Artificial Plants

Artificial plant optimization algorithm (APOA) is a novel evolutionarycomputation inspired by plant growing process which contains photosynthesisoperator, phototropism operator and apical dominance operator. Due to the wideliving area and slow growing speed, animal is more suit than animals in nature.This biological background means that APOA is fit for solving multi-modal,high-dimension problems. In this paper, we focus to phototropism operator, andtwo modifications are designed to improve the performance. The maincontributions are listed as follows:(1) In the standard version, the influence of environment is omitted inphototropism operator. However, it conflicts with natural phenomenon. As weknown, the growing speed is affected by environment significantly. Therefore, anew strategy–double shrinkage factors are introduced in phototropism operatorto reflect two type of environments, and we call it APOA with double shrinkagefactors (APOA-DS). Furthermore, convergence analysis is provided by Markovchain. Simulation results show it is superior to other five algorithms formulti-modal high-dimension problems.(2) APOA-DS incorporates the environment influence into phototropismoperator, with this manner, it can easily find one range in which global optimumis fallen. However, the local search capability is very poor. To overcome thisproblem, one local random mutation operator is designed and is introduced intomethodology of APOA-DS. To avoid the confusion, we call it APOA withdynamic local search strategy (APOA-DLS). Simulation results showAPOA-DLS is superior to APOA-DS.(3) To further testify the performance of two proposed algorithms, we applyAPOA-DS and APOA-DLS to solve the optimal coverage configuration ofwireless sensor networks. Simulation results show the standard version of APOAis the worst, while APOA-DLS is superior than APOA-DS.