Computer foundation laboratory is an important teaching platform for computer teaching in the university,and is also key measure for undergraduates archiving the ability of computer application. The performance evaluation on computer foundation laboratory is analyzed in order to provide the references to its management,and then to improve management efficiency. A performance evaluation model is built with projection pursuit technique and performance evaluation index. The model consists of laboratory comprehensive efficiency,laboratory teaching effect,laboratory social service effects and laboratory comprehensive management. The performance evaluation model based on projection pursuit is verified to be effective according to an example,and its result is trustworthy.

Internet of things is a strategic emerging industry and is paid close attention and promoted by all country.It analyzes the urgency of constructing laboratory of internet of things specialty,and gives the goals and the principles of laboratory construction according to specialty training goals.The construction framework is designed based-on foundation laboratory,project training room,and comprehensive project demonstration laboratory.

为提高组搜索优化算法求解多维函数优化问题的性能,提出一种结合逐维搜索、Metropolis准则、反方向视角和禁忌表策略的改进组搜索优化算法.逐维搜索策略逐维更新并评价成员位置,在每一维,更新的值与其他维组成候选位置,使用模拟退火的Metropolis准则来决定是否接受候选位置.反方向视角策略使成员按一定的概率做反方向搜索,禁忌表策略避免生产者始终保持不变.这些策略能更好地平衡算法的集中性和多样性.在典型测试函数上进行了仿真,结果表明改进策略是有效的,提高了组搜索算法求解多维函数优化问题的全局寻优能力和收敛速度.

This paper presents a modified group search optimizer algorithm for high dimensional function optimization, which is based on levy flight strategy, self-adaptive joining strategy, and chaotic mutation strategy. The levy flight strategy is employed for the producer to simplify the computation and improve efficiency in the exploring space. The self-adaptive joining strategy is used for the scroungers walking towards the producer to promote convergence speed. The chaotic mutation strategy is designed for the rangers to strengthen diversification. Using those strategies, the modified algorithm can get better balance between intensification and diversification. The simulation experiments, which were carried on benchmark functions, show that those strategies are effective, and they improve the global optimization ability and convergence speed of modified group search optimizer for high dimensional function optimization.

In order to determine the optimal flocculants and the corresponding dosage,the effects of four flocculants(including aluminum sulfate,polyaluminium chloride,ferrous sulfate and calcium oxide) were studied for removing cadmium,lead,zinc from the suspension containing fine soil particles separated by soil washing process.The concentrations of cadmium,lead and zinc in the suspension were decreased firstly and then increased with the increasing dosage of aluminum sulfate.The concentrations of these three heavy metals were rapidly decreased with the increasing dosage of polyaluminium chloride and calcium oxide.However,when the dosage of ferrous sulfate increased,the concentrations of lead and zinc were decreased,and the concentrations of cadmium rapidly were decreased and then slightly increased.By comparison,the best flocculation dosage of these four flocculants for the flocculation and precipitation of the particles in the suspension were 0.25,0.04,1.0 and 0.3 g/L,respectively.When the suspensions treated with the optimal dosage of each flocculants,the concentrations of these three heavy metals were all below the irrigation water quality standards of China.Using the optimal dosage,the cost of flocculant for treating the same volume of suspensions were ferrous sulfate > aluminum sulfate > polyaluminium chloride > calcium oxide.In conclusion,calcium oxide was the optimal flocculants for treating the suspension containing fine particles,and the corresponding optimal pH range was 10 to 11.

The particle swarm optimization clustering algorithm was applied to classify fireproof trees,which were 37 species of coniferous and broad-leaf trees with 10 fireproof indexes in Fujian province.The fireproof trees were divided into six types.The result of classification showed it achieved a perfect effect,and accorded with production practice.Comparing with the ant colony clustering,the result showed that better fitness value,the longer distance between clusters,and the shorter distance inside clusters were archived when the particle swarm optimization clustering was applied to analyze the fireproof trees.The algorithm is convenient to be applied,and can propose a new method for the related research on forestry science.

The parameters of the site index curve model were solved by using particle swarm optimization (PSO) algorithm,with the minimum error in estimation of dominant height and site index as the target function.An application example is introduced to test the proposed method,and the estimation result is compared between PSO and immune algorithm.Comparison result shows that the parameters solved by PSO can decrease the overall error,increase the precision,improve the fitting effect of the site index curve model,thus increase the estimation precision of young forest.This research is hoped to provide new idea for parameter solving of grow model in forest management and related research,and expand the application of PSO in forestry science as well.

The forest landscape pattern of Longqishan National Nature Reserve was analyzed by using landscape indices,such as patch area,patch perimeter,number of patch,fractal dimension,separation index,dominance index,and so on,which were calculated under the landscape structure quantitative software.The results showed that broad-leaved,pine and bamboo were main landscape styles of the reserve.They also revealed the distribution of the areas,the perimeters,and the number of patch were very unevenly,and presented that the broad-leaved landscape was the most complicated,and that its fragmentation degree and patch separation degree were fairly lower,but non-wood landscape and economic landscape were higher.Meanwhile,they indicated that the reserve had a high predominance,uneven landscape distribution and a low diversity on the whole.

The precision of ideal site index curve model mainly depends on the solved parameters of the fitting equation. To archive high precision of model, a particle swarm optimization algorithm with iterative improvement strategy was proposed to solve parameters of the model. The improved algorithm makes each particle update it's current velocity and position dimension by dimension. The result shows the parameters solved using particle swarm optimization with or without iterative improvement strategy make the model to be small overall error, high precision, ideal of fitting effect, scientific, and reasonable. It also indicates that particle swarm optimization with iterative improvement strategy is better than particle swarm optimization without iterative improvement strategy on the performance with the same conditions. The study provides a new way for solving parameter of growth model in forest management, and for the related research. It also enriches not only optimization technology about stand management, but also the application domain of particle swarm optimization algorithm. It can be predicted that particle swarm optimization algorithm will be the broad application prospect in the forestry production and scientific research.

Based on projection index function,projection pursuit cluster analyis is a process to analyze optimization problems under certain constrain conditions.This paper employs Particle Swarm Optimization(PSO) to solve projection index function,and constructs a projection pursuit cluster model to evaluate forest carrying capacity.Simulation results show that this PSO-based model is simpler and easier to realize with less parameters,compared with GAbased model.Moreover,the proposed model can give optimal solutions in application,which verify its practical significance in forest carrying capacity evaluation and other regional sustainable development research.

A maintainability prediction model was built by using the principle of support vector regression.The predictors were defined as object-oriented software metrics,and the maintainability was measured as the number of changes made to code during a maintenance period.To evaluate the performance of model,the artificial neural networks model was also built.Simulated SVM model under the environment of R software had a better performance of predicting the maintainability,and was superior to ANN model by visual analysis of error and RMSE.

This paper carries on evaluating forest landscape by applying projection pursuit based-on genetic algorithm,which converts more indexes of forest landscape into a projection value.The method is to get the best projection direction by applying each level value of evaluation index,and to set section according to comprehensive projection value of each index,then,to draw an evaluation conclusion by comparing the comprehensive projection value of object with sections.Taking forest landscape of Mangdang mountain nature reserve in Fujian Province as a sample,the result of evaluation shows it's feasible for applying projection pursuit to forest landscape evaluation.It also predicts that it has an import application value for applying projection pursuit based-on genetic algorithm to forest landscape,and establishes foundation for extensive application in sustainable development with the improvement on projection pursuit.

The forest resource assets evaluation system is an important tool for evaluation,but this system can not fit for common requirement of consultation service for users.The forest resource assets evaluation Web service system is studied and designed,which includes Web application system,Web service and its register,the solution scheme was implemented by using C# language based on Net.

This paper carried on constructing forest landscape evaluation model based on radical basis function(RBF) network,by selecting level standard data of evaluation index as train sample.The method was to take Mangdang Mountain Nature Reserve of Fujian Province,as research object,and to survey on the scene as evaluation sample.The result of evaluation on forest landscape is 2nd level the same as others.It has proved that it is feasible to apply RBF network to evaluate forest landscape.

The ant colony algorithm cluster analysis was applied to classify fireproof trees, which were 37 species of coniferous and broad-leaf trees on Fujian province. They were divided into six classes, according to ten fireproof performance. Realized under the Matlab environment, the result of classification showed it achieved a perfect effect, and accorded with production practice. Thus, it verifies an application perfect about applying ant colony algorithm cluster analysis to classification of fireproof trees, and proposes a new ideal and method for modeling of forest ecosystem.

Cyclobalanopsis chungii is unique valuable timber species in China.Based on investigation of Cyclobalanopsis chungii community of Minqing,Fujian province,the feature and diversity of Cyclobalanopsis chungii community were studied in this paper.The results showed that in tree layer,Cyclobalanopsis chungii were the most important tree species;In shrub layer,Castanopsis fargesii,Ardisia punctata,Adinandra millettii and Phoebe bourn had predominant status in Cyclobalanopsis chungii community;In herb layer,Allyxia sinensis were the most important species; The order of species diversity was shrub layer,tree layer and herb layer.

In this paper,based on the theory of ecological which using the qualitative analysis and landscape evaluate index sign system evaluated the forestry landscape of Fujian Mandangshan Nature Reserve.Such indices as naturalness,rarity,diversity,representativeness,scientific researching were chose to evaluate the forestry landscape of the nature reserve,then measured off the results some classifications.It suggested that the forestry landscape of the nature reserve was achieved the second degree and had relatively well value of the ecotorism.Besides,the reserve's problems were analyzed and corresponding proposals were put forward.

本文介绍高莫雷割平面法及其求解森林经理中的人员和设备分配等整型规划问题.通过分析高莫雷割平面法原理,设计算法流程图,并编制程序.在具体实例上对算法进行了测试,结果与现实相符合.

With the cognition of development of the forest certification all over the world,the authors intend to outline the importance of forest certification of the forestry industry in China and to establish principles and correlation standards by using for conferences to foreign forestry industry system and according to facts of forestry industry in China.Therefore,it can be concluded 8 principles and 40 standards and give some advices for forest certification at present which are favorable to promote the development of forest certification and improve the forest sustainable management for state-owned forest center.