线性规划、模拟退火和遗传算法在杉木人工林可持续经营中的应用和比较

以杉木人工林为例 ,介绍了线性规划、模拟退火和遗传算法在编制森林经营方案过程中的应用和比较 ,同时介绍了一种通用的可以处理绝大多数的森林经营模型。采用Hui(1997)的生长和间伐模型来模拟林分的生长、间伐和发展过程。线性规划、模拟退火和遗传算法三者适用于不同的场合。当约束条件都比较宽松时 ,线性规划也有可能得出整数解 ,但不一定就能够避免林分分割经营 (整数解 )。要获得整数解 ,尤其是在林分数目很大的时候 ,可以采用模拟退火和遗传算法。如果允许林分分割 ,线性规划的结果一般可以获得最好的目标方程值。

The Application and Comparison of Linear Programming, Simulated Annealingand Genetic Algorithm in the Sustainable Management of Cunninghamialanceolata Plantations

The methods of Simulated Annealing and Genetic Algorithm were introduced by using an example of %Cunninghamia lanceolata% plantation. Their application in sustainable forest management was compared with Linear Programming by the common model and the same data set. Basic growth and yield information were provided using a model developed by Hui (1997). Linear Programming will establish an optimal solution if it exists, but stand splitting cannot be avoided. Simulated Annealing and Genetic Algorithm converge to an optimum (or near_optimum) resulting in an integer solution, but the proper parameter setting is crucial. If stand splitting is allowed, and the constraints are sufficiently tight, the Linear Programming solution can be expected to be better than that of Simulated Annealing and Genetic Algorithm.