优化问题

It is related to the variational inequality problem, bilinear programming, nonlinear equation. So there is actual meaning and theories value to study how to solve it.In recent years, many algorithms for solving the generalized linear complementarity problem are put forward. But they basically belong to the traditional iterative methods, and may not be efficient since the computing time required for a solution is greatly dependent on the dimension, structure of the problem and the complexity of the algorithm used. Different from the traditional algorithms, the neural network has many advantages on calculations and real-time applications for its inherent massive parallelism and electric circuit implementation.

近年来，人们对广义线性互补问题提出了许多算法，但这些方法基本上属于传统的迭代法，由于它们的计算时间极大地依赖于问题的规模、维数以及所使用的算法，因此很难满足实时并行的要求，与传统数值方法不同，由于内在的并行分布处理信息的特点及电路实现的潜能，神经网络有着许多计算上的优势和实时性的应用，自Hopfield提出著名的人工神经网络模型-Hopfield神经网络，并将其成功应用于优化问题后，用神经网络求解优化问题得到了相当深入的研究，并取得了许多重要的成果。

It is difficult for a variety of traditionary methods based grads information to solve the constrained optimization problem with non-differential coefficient,multi-modal,nonconve and non-linear functions.

在现实生活中许多实际问题都可以转化为约束优化问题，并且实际问题通常都很复杂，其函数形态各具特色，传统基于梯度信息的各种求解策略对于具有不可微、多峰及非凸的非线性函数约束优化问题很难凑效。

The thesis studies the difference between GWSs composition and traditional workflow applications, proposes the implement mechanism and technical architecture of GWSs composition and defines the reference model of GWSs composition platform, which provide holistic framework and top-level guidance for studying dynamic GWSs composition technology, designing GWSs composition platform, and implementing representative composite service applications.2、Theories and methods of composite process modeling and process analysis. This thesis proposes a meta-model of Geo-spatial Web services composition process defining from two aspects, i.e. process and resource. According to the process definition meta-model, we propose a new Geo-spatial Web services composition Process/Resource model based on extended WF-net, which effectively resolves the problems on uncertain activity option and dynamic variation of service in composite process model. A notion with regard to soundness of process model is proposed from three aspects, i.e. the structural correctness of process model, validity of data link and validity of resource. The notion of soundness definitely gives a dividing line on correctness of composite GWS process. This thesis analyzes sufficient and essential terms of soundness, and studies compositional property of WSCP/R-net models which provides an effective method to construct a soundness WSCP/R-net model. According to WSCP/R-net, the thesis proposes composition algebra and studies the properties of it. Based on the composition algebra, 6 kinds of composite process reduction rules are induced to optimize the process.3、Geo-spatial Web service QoS model and its application architecture. This thesis proposes an extendable GWS QoS model from three aspects, i.e. Geo-spatial Web Service itself, networking conditions and service consumer which can be used to distinguish GWSs with overlapping or identical functionality. We work over the QoS evaluation methods of composite GWS process which can be used to guide the execution, monitor and service selection of composition process. Aming at some disadvantages in traditional Web service discovery model for its lacking of QoS supporting, this thesis proposes a new service discovery model SDMQ supporting QoS constraints.4、Technologies of composite service executing. SSPL, a new composite service process language, is defined to overcome the disadvantages of existing composite services process languages which can not adapt to the dynamic variation of Geo-spatial Web Services. An algorithm is described to translate WSCP/R-net model into SSPL. The thesis studies the model of dynamic services selection with QoS global optimization, presenting an algorithm GODSS to resolve dynamic services selection with QoS global optimization in GWSs Composition.5、Design and implement experimental system and representative applications. According to the research findings presented above, we design and implement an experimental system and construct two representative applications to show that our achievements are effective and practical.

针对当前空间信息服务聚合研究中存在的若干不足，本文重点在基于工作流的空间信息服务聚合框架、空间信息服务聚合流程建模和分析技术、空间信息服务QoS模型及应用体系以及空间信息服务聚合流程执行等几个方面进行深入研究，主要工作和创新点包括：1、基于工作流的空间信息服务动态聚合框架：研究了空间信息服务动态聚合和传统工作流应用的不同，在此基础上提出了基于工作流技术的空间信息服务动态聚合实现机制、技术体系和参考模型，为研究空间信息服务聚合的关键技术、开发服务聚合支撑平台以及在此基础上构造特定的空间信息服务聚合应用提供了总体框架和顶层指导。2、空间信息服务聚合流程建模和分析技术：首先基于动态服务聚合流程建模的需求和空间信息服务动态变化的应用实际，从服务聚合的过程维和资源维出发定义了服务聚合流程定义元模型；与之相对应，对基本的WF-net进行扩展，提出服务聚合流程/资源网作为空间信息服务聚合流程的形式化描述模型，有效解决了现有的基于基本Petri网和工作流网的服务聚合建模方法所不能解决的不确定路径选择和服务的动态变化性问题；从流程结构正确性、数据依赖有效性和资源实现有效性三个方面提出了WSCP/R-net健壮性的概念，明确界定了聚合流程正确性的标准，并对WSCP/R-net模型的健壮性分析方法和WSCP/R-net模型的组合特性进行了研究；提出了空间信息服务聚合代数算子并对其性质进行了研究，在此基础上提出了6种聚合流程约简规则，从而可达到优化聚合流程、提高聚合流程执行效率的目的。3、空间信息服务QoS模型和应用体系：从服务本身、网络环境和服务消费者三个层次出发，提出了一个可扩展的空间信息服务QoS模型GSQM，实现了对空间信息服务的度量和评价，并对GSQM不同质量要素信息的客观、公正采集方法进行了研究；研究了空间信息服务聚合流程QoS评价方法，有效支持了空间信息服务聚合流程的执行、监控以及服务选择等操作；针对目前的服务发现模型仅支持服务功能性描述、不能有效满足空间信息服务应用需求的现状，提出了一种新的支持QoS约束的服务发现模型SDMQ，并对模型的实现框架进行了研究。4、空间信息服务聚合流程执行技术：针对现有的服务聚合流程描述语言不能有效满足空间信息服务动态聚合流程描述的特点，基于BPEL4WS提出了一种新的空间信息服务聚合流程描述语言SSPL；研究了满足健壮性要求的WSCP/R-net模型向SSPL的转换算法，在此基础上可以满足用户无显式编码来实现空间信息服务聚合应用的目的；研究了服务动态选择QoS全局优化模型，并在此基础上提出了一种解决服务动态选择QoS全局优化问题的实现算法GODSS.5、基于论文研究成果，设计实现了"空间信息服务聚合实验系统"，并构建了"矿产资源评价"和"城市消防应急响应"两个典型应用案例，对论文所述模型、方法的可行性和有效性进行了验证。

Wanka studied two kinds of dualproblem in finite dimensional space, namely, an extended Fenchel type and Fenchel-Lagrange dual based on the conjugate theorem in convex optimization problem.

Wanka利用有限维空间中凸优化问题的共轭理论，研究了两类对偶问题，即广义Fenchel对偶问题和Fenchel-Lagrange对偶问题，后者是经典Fenchel和Lagrange对偶问题的组合，二者都是在扰动理论基础上产生的，还提出了一个约束条件保证其凸优化问题中强对偶成立。

W anka studied two kinds of dualproblem in finite dimensional space,namely,an extended Fenchel type and FenchelˉLagrange dual based on the conjugate theorem in convex optimization problem.The latter is a"combination"of the classical Fenchel dual and Lagrange dual,which are based on perturbation theory.Then they introduced a constraint qualification whose fulfillment is sufficient in order to guarantee strong duality.

W anka利用有限维空间中凸优化问题的共轭理论，研究了两类对偶问题，即广义Fenchel对偶问题和FenchelˉLagrange对偶问题，后者是经典Fenchel和Lagrange对偶问题的组合，二者都是在扰动理论基础上产生的，还提出了一个约束条件保证其凸优化问题中强对偶成立。

The layout scheme design of spacecraft vessel belongs to 3D packing problem with behavioral constraints, and it is also a complex engineering system problems with inscrutability, uncertainty, high nonlinearity, quality and quantity. It is almost impossible to solve it completely by human or computer alone.

航天器舱布局方案优化设计属带性能约束的三维布局优化问题，属具有不确定性、高度非线性、既有定性又有定量问题的复杂工程系统问题，单纯依靠人或计算机算法去解决都是非常困难的。1991年Lenat、Feigenbaum提出的人机合作(Man-Machine Synergy)和同年钱学森、戴汝为等提出的人机结合(Human-Machine Cooperation)思想和方法论，被认为是21世纪解决复杂工程问题的重要途径、自动化设计的重要发展方向，但其可操作性问题一直未很好解决。

Distinct from those methods, this paper's main idea is to reformulate CPs as an unconstrained minimization problem with the form of least squares of a system of overdetermined equations being equivalent to the CPs.

与这些方法不同，本文的基本思想是借助互补问题与某个超定方程组同解的事实，利用该方程组的最小二乘形式将原问题直接转化为无约束优化问题，这样不仅把问题转化为光滑优化问题，而且避免了引入光滑参数导致的不便。

Based on Pontryagin maximum principle,the lunar soft landing problem is transformed into a two-point boundary value problem in mathematics.

首先，从庞德里亚金极大值原理出发，将有限推力作用下月球软着陆问题转化为数学上的两点边值问题；在考虑边界条件及横截条件的前提下，将该两点边值问题转化为针对共轭变量初值和末时刻的优化问题；然后应用非线性规划方法求解所形成的参数优化问题。

Based on Pontryagin maximum principle, the lunar soft landing problem is transformed into a two-point boundary value problem in mathematics. In consideration of the bound condition and transversality condition, the resulted two-point boundary value problem is converted into a parameter optimization problem aiming at the initial values of conjugate variables and the terminal time, then it is solved by the nonlinear programming.

首先，从庞德里亚金极大值原理出发，将有限推力作用下月球软着陆问题转化为数学上的两点边值问题；在考虑边界条件及横截条件的前提下，将该两点边值问题转化为针对共轭变量初值和末时刻的优化问题；然后应用非线性规划方法求解所形成的参数优化问题。

The modified particle swarm optimization is utilized to solve the proposed reactive power optimization model; and combining with non–stationary multi-stage assignment penalty function, the reactive power optimization problem is transferred into a problem to solve extreme-value without constraints, thus the calculation accuracy and global convergence ability of MPSO are effectively improved, and power grid company can achieve maximal economic benefit by means of the proposed algorithm.

应用改进粒子群优化算法求解该无功优化模型，并结合动态调整罚函数法将无功优化问题转化成无约束求极值问题，从而有效地提高了改进粒子群优化算法的全局收敛能力和计算精度，使电网公司取得了最大经济效益。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。