Font Size: a A A

The Research On Dynamic Multi-objectives Optimization Of Forest Space Based On Multi-swarm PSO Algorithm

Posted on:2018-09-27Degree:MasterType:Thesis
Country:ChinaCandidate:J HuangFull Text:PDF
GTID:2348330515459125Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
There are many complicated dynamic multi-objective optimization problems in practical engineering field.At least one of the objective functions,constraints,or definitions of dynamic multi-objective optimization problems varies over time,so that the optimal solution will change over time.Multi-swarm particle swarm optimization algorithm is an algorithm to solve the multi-objective optimization problem,which is based on the parallel search of multiple sub groups.The algorithm is feasible to solve the static multi-objective optimization problem.But it is lack of the detection of the change of the environment when solving the dynamic multi-objective optimization problem.Therefore,the improved multi-swarm particle swarm optimization algorithm is proposed which introduces an environment detection operator.Then a dynamic multi-objective optimization model of forest spatial based on the improved MPSO algorithm is build.The main contents of this paper are as follows:1.A new algorithm for solving dynamic multi-objective optimization problem is proposed and the performance of the improved algorithm is tested.The fitness function of dynamic multi-objective optimization problem is the basis of environmental change.The mean value of the Euclidean distance difference between the new and the old fitness functions is as the environmental detection operator ’ by randomly selecting 20%individuals in the particle population.Then the changes of environment is judged by comparing the size of the actual problems of the preset threshold value θ and ε.In order to improve the global and local search ability of the algorithm,the performance of PSO algorithm is analyzed and compared under five kinds of inertia weights.The performance of the improved algorithm are tested by four typical dynamic multi-objective optimization problems.By comparing with the PDMIOA and dDMS-MOPSO algorithm,the improved algorithm in this paper is better in convergence,distribution and tracking performance.2.The dynamic multi-objective optimization model of forest spatial structure is constructed.According to the single tree growth model of tree growth change along with the time,forest mingling and competition index and angle scale and spatial index and canopy density index and open comparison and neighborhood comparison and individual volume and health index from the three aspects of traditional forest structure and forest spatial structure and vertical structure are selected as the goal function to establish a dynamic forest spatial structure optimization model.And the model is solved by the improved algorithm in this paper.3.The dynamic multi-objective optimization model of forest spatial structure is verified.In view of the actual stand,three environmental variables were set up for a period of three years.And the effect of stand adjustment on the stand structure was considered as the standard for the two experiments.Firstly,the output results of the two experiments were analyzed and the corresponding adjustment strategy of the forest was put forward.Then through the comparison before and after the adjustment of stand heterogeneity index and mixed degree and competition index and health index and spatial density index and open comparison,found that the stand two experiments under the adjusted index are better than the unadjusted index and stand structure obviously improved.Therefore,the dynamic forest spatial structure optimization model proposed in this paper is reliable and can provide technical support for the development of intelligent forestry.
Keywords/Search Tags:Dynamic multi-objective, Multi-swarm particle swarm optimization, Pareto optimal solution set, spatial structure optimization, tracking error
PDF Full Text Request
Related items