【作者】 万喜喜；

【导师】 唐代生；

【作者基本信息】 中南林业科技大学 ， 森林经理学， 2009， 硕士

【摘要】 复杂性科学是用以研究复杂系统和复杂性的一门学科,它被有些科学家誉为是“21世纪的科学”。目前,关于复杂性的研究受到了世界各国科学家们的广泛关注。自复杂性问题的提出至今,已经取得了不少研究成果,而其中最有优越代表性的系统理论数CAS(复杂适应系统)理论,已被广泛地应用于生态、社会、经济、管理、军事等领域复杂性问题的研究。本文首先运用CAS理论从聚集、多样性、非线性、流、标识、内部模型和积木七个方面分析了森林生态系统的复杂适应性,论述了森林生态系统是一类复杂适应系统,然后介绍了CAS理论的相关成果,选取与CAS理论有紧密联系的遗传算法作为研究对象。作为一种智能优化算法,遗传算法的内在并行机制及全局优化的特点适合于多目标规划问题的解决。因此本文针对森林收获调整这种大型复杂系统的决策问题,引入了Pareto多目标遗传算法求解森林收获调整多目标规划模型,讨论了群体初始化、评价函数、选择算子、交叉算子、变异算子以及迁移算子等关键步骤。本文严格遵循视调整最终目标是实现森林资源永续利用的思想,以江西景德镇枫树山林场为实际算例,以一般用材林的四个主要优势树种(马尾松、杉木、湿地松和阔叶林)作为调整对象,运用欧式距离和平衡率两个数量指标对森林收获调整的效果进行了评定,评定结果：4个树种龄级结构的平衡率都有较大的提高,欧氏距离都缩小了90%以上,到调整期末森林资源的龄级结构基本上达到了完全调整林状态,收获调整的效果明显,取得了预期的目的,有利于森林的永续经营。调整结果表明,在森林收获调整中引入遗传算法是可行、有效的,而且能为决策者提供满意解,这是运用CAS理论解决森林系统相关问题的一个重要方面,CAS理论对我们理解森林的复杂性,解决进行森林经营活动遇到的实际问题有重要意义。更多还原

【Abstract】 Science of complexity is a kind of subject that is used to study complex system and complicacy, it is praised by some scientists as "the science of the 21st century". At present, scientists from all over the world pay wide attention on the research that about the complexity. Since the bringing forward of complexity to present, has made a lot of research results, and the most superiorly representativeness theory of system is the CAS (Complexity Adaptive System) theory, it has been widely used in ecological, social, economic, management, military and other areas of research about complex problem.This paper firstly uses the CAS theory to analyze the complex adaptability of forestry ecological system from the aggregation, diversity, non-linearity, flows, tagging, internal model and building block, discoursing upon the forestry ecological system is a class of complexity adaptive system, and then introduce the relevance of the results of CAS theory, select genetic algorithms that closely related to CAS theory as the object studied.As an intelligent optimization algorithm, the genetic algorithm’s parallel mechanism and the intrinsic characteristics of global optimization is suitable for multi-objective programming problems. Therefore this paper aim at the Forest Harvesting Regulation such kind of large and complex system of decision-making problem, the Pareto multi-objective genetic algorithm is introduced to solve the multi-objective programming model of Forest Harvesting Regulation, the key steps of group initialization, evaluation function, selection operator, crossover operator, mutation operator and the transfer operator are discussed.This article strictly follows the idea that regards the achievement of sustainable use of forest resources as the ultimate goal, taking Jiangxi Jingdezhen Maple Forest Farm for a practical example, taking the four dominant tree species of general timber stands (Pinus massoniana Lamb, Cunninghamia Lanceolata, Pinus elliottii and large tall broad- leave tree) as the controlled member, the utilization of Euclidean distance and balance ratio two quantitative indices has carried on the evaluation to the effect of Forest Harvesting Regulation, the evaluation result:the age-grade structure’s balance ratios of four dominant tree species have the big enhancement, Euclidean distance is away from reduced above 90%, basically forest resources’ age-grade structures have been adjusted to the forest condition on the whole at the end of the adjustment period, the harvest adjustment effect is obvious, has obtained the anticipated goal, was propitious to sustainable operation of forest. Adjustment results show that introducing the genetic algorithm into Forest Harvesting Regulation is feasible and effective, and can provide satisfactory solution for decision-makers, this is an important aspect of the use of CAS theory to solve the problem of forest-related, CAS theory has significance for us in understanding of the complexity of forests and solving the practical problems that encountered in carring out of forest management activities.更多还原