Abstract:

Although intellectualized afforestation design is the important technical guarantee in efficient culture of industrial plantation,appropriate methods are still lacking in multi-species afforestation.Genetic algorithm presents the intellectualized characteristics in two-species afforestation due to the use of 0-1 coding.An improvement of genetic algorithm was put forward in this paper,which optimized the objects by four genetic manipulations of copy,exchange,inversion and cross mutation with its genetic coding expressing as 1,2,…,m (numbers of tree species).In the aspects of funds,plans,seedling varieties and numbers,the mathematical model and precise solution of optimal afforestation technique were established in industrial plantation construction.The improved genetic algorithm was applied in the afforestation technical design of 50 sub-compartments of Fujian Yong an Forestry (Group) Joint-Stock Co.,Ltd.Based on the site index and growth model of industrial plantation of Cunninghamia lanceolata,Pinus massoniana and Eucalyptus spp,the optimal design was achieved by the application of improved genetic algorithm.For 129.66 hm²of C.lanceolata afforestation,454 900 seedlings were needed and the cumulative timber yield would reach 20 100 m³ in twenty years;for 80.04 hm²of Eucalyptus spp.afforestation,149 990 seedlings were needed and the cumulative timber yield would reach 14 440 m³ in seven years,three-batch timber yield would reach 43 320 m³ after 20 years;for 61.83 hm²of P.massoniana afforestation,276 720 seedlings were needed and the cumulative timber yield would reach 7 330 m³ in twenty years.Guided by the optimal afforestation technical design,the total afforestation area of industrial plantation reached 271.53 hm²,total investment reached 1.448 284 million and total profits would reach 41.527 8 million after 20 years.The expected effect was obtained and therefore multi-species afforestation technique could be applied and promoted in forestry practice.

Key words: genetic algorithm, afforestation technique, multi-species design