Study On Optimization Of Stand Spatial Structure Based On Weighted Delaunay And Immune Algorithm

The spatial structure of forests reflects the interaction between forests,which is of great significance to the long-term stable development of forests.In this study,Hunan Lutou Forest Farm and simulated sample plot were taken as the research object.Considering the influence of tree size on its interaction with surrounding trees,the DBH,tree height and crown width of trees are generated into comprehensive weights,and the distance between the center tree and adjacent trees in the spatial structure unit determined by Delaunay triangulation network is weighted.The physical coordinates of trees were transformed into logical coordinates,the size attributes of trees were transformed into spatial attributes,and the topological relationships of interactions among trees were changed.The main research contents and results are as follows:(1)Considering trees of different sizes as individuals with different competitive status,a weighting method based on Delaunay triangulation model was proposed to generate weighted spatial structure units.The number of the adjacent trees before and after weighting was 5.By weighting,26.99%and 23.62%of the adjacent trees in natural and simulated plots remained unchanged,and the range of changes in the number of adjacent trees was small.In the weighted tree spatial structure unit,the average distance increment between the central tree and its adjacent trees decreases with the increase of the comprehensive weight of the central tree,which indicates that the tree topological relationship obtained by the weighted method can more accurately characterize the strength of the interaction between the trees.(2)The spatial structure indexes were calculated based on weighted and non-weighted spatial structure uints,and their applicability were discussed.The normal distribution of spatial structure index before and after weighting was tested by graphic method and computational method.Seven spatial structure indexes before weighting were consistent with normal distribution under the conditions of verification by two methods.After weighting,the neighborhood comparison,competition index,open degree and spatial density index tested by different methods have some deviations.Considering comprehensively,all spatial structure indices before and after weighting basically conform to normal distribution.Spearman correlation analysis of pre-weighted index and post-Weighted index showed that there were significant correlations among size ratio,forest layer index,mingling and open degree,indicating that the above weighted spatial structure indexes were effective indexes for describing the spatial structure of stand;there was no significant correlation in both competition index and the uniform angle index,but they could describe the spatial structure of stand more accurately.The homogeneity index in all plots except N5 was significantly positively correlated,and the paired sample T test showed that there were significant differences between the homogeneity index in all plots except N1 and N2.It showed that the weighted homogeneity index was still applicable to the spatial structure evaluation of forest,and the result analysis is more scientific and reasonable.(3)A weighted spatial structure optimization thinning model was constructed with 7 weighted indices such as size ratio and forest layer index as the objective of spatial structure optimization and with the constraints of tree size diversity,diameter class diversity and thinning amount.Artificial immune algorithm was used to solve and formulate thinning scheme.The N4 sample plot was selected for experimental analysis,and the results further verified the reliability of the weighted spatial structure thinning optimization model and the feasibility of solving the optimization model by immune algorithm.In this paper,the applicability of weighted spatial structure index and homogeneity index were analyzed.On this basis,the optimization model of thinning of forest spatial structure was improved,which provides theoretical basis for rational optimization of forest spatial structure and provides new ideas and approaches for healthy forest management.