针对浅层学习模型对风速预测存在较大误差的缺陷,提出一种基于小波变换和深度信念网络(wavelet based deep belief netw ork,WDBN)的风速预测模型。首先利用小波变换将原始风速序列分解成不同的频率序列;然后针对各频率序列,根据逐层训...针对浅层学习模型对风速预测存在较大误差的缺陷,提出一种基于小波变换和深度信念网络(wavelet based deep belief netw ork,WDBN)的风速预测模型。首先利用小波变换将原始风速序列分解成不同的频率序列;然后针对各频率序列,根据逐层训练法则设计深度信念网络模型;最后采用小波反变换对不同频率序列重构,得出最终的风速预测结果。选用某风电场2013年1月和7月的数据对WDBN模型的风速预测进行仿真分析,并与自回归滑动平均法、反向传播神经网络法、Morlet小波神经网络法的预测结果进行对比,结果表明WDBN模型可以更好地学习风速所具有的非线性和非平稳特征,具有较高的预测精度。展开更多
In a composite-step approach, a step Sk is computed as the sum of two components Uk and hk. The normal component Vk, which is called the vertical step, aims to improve the linearized feasibility, while the tangential ...In a composite-step approach, a step Sk is computed as the sum of two components Uk and hk. The normal component Vk, which is called the vertical step, aims to improve the linearized feasibility, while the tangential component hk, which is also called horizontal step, concentrates on reducing a model of the merit functions. As a filter method, it reduces both the infeasibility and the objective function. This is the same property of these two methods. In this paper, one concerns the composite-step like filter approach. That is, a step is tangential component hk if the infeasibility is reduced. Or else, Sk is a composite step composed of normal component Uk, and tangential component hk.展开更多
In this paper, we account for this subject: how to de sign a pattern, it can track the state of the equipment of some organizations su ch as enterprise, organ, laboratory, school etc. We present an analysis pattern, w...In this paper, we account for this subject: how to de sign a pattern, it can track the state of the equipment of some organizations su ch as enterprise, organ, laboratory, school etc. We present an analysis pattern, which describes the whole procedure of managing the equipment and record the us ing of the equipment. It not only can track the quantity and location of the equ ipment of the whole organization, the more important is it can update the state of the equipment at real-time automatically. First,we design the static diagram(i.e. using UML class diagram to describe the basic state of the equipment). Then we consider its dynamic aspect, i.e., how th e state of the equipment to get changed according to the time. We use UML sequen ce diagram and state diagram to respectively describe the procedure of after pur chasing, transferring and discarding as useless of the equipment. Obviously, the manager can update the quantity and location of the equipment automatic ally. We character this pattern from the following five aspects: Problem: How the enterprise, organ, laboratory and school to track the quantity and location of the equipment. Circumstance: In some organizations, especially, in the manufactory or laborator y, when the number of the quantity and type is large or the distribution of the equipment is dispersed, they want to be able to track the quantity and location of the equipment. Forces: First, it is possible to the equipment be transferred or be discarded, n o matter when and where, the organization must be able to track the factual quan tity and location.Second, the solution must describe a basic semantic unit, that is, the solution must simple enough to apply it to various of circumstance, whi ch is the base of reusing. Third, the solution must include the interpret of the factual document. Solution: This part, we start with the basic demands, first using UML class diag ram to describe the basic pattern, which is an atomic pattern. Then using UML se quence diagram and state diagram to respectively de展开更多
现有未知突发信号检测算法是基于噪声加单一突发信号的简单假设的,在实际复杂信号环境会产生大量虚警而失效。针对实际非合作突发通信信号的检测环境除噪声外还包含多个连续信号和一些短突发干扰信号,建立了复杂信号环境模型,提出了适...现有未知突发信号检测算法是基于噪声加单一突发信号的简单假设的,在实际复杂信号环境会产生大量虚警而失效。针对实际非合作突发通信信号的检测环境除噪声外还包含多个连续信号和一些短突发干扰信号,建立了复杂信号环境模型,提出了适用于此环境的基于短时傅里叶变换(short time Fourier transform,STFT)的时序检测器。该检测器利用突发通信信号时间上短持续的特点剔除连续信号和短突发干扰造成的虚警。对该检测器的检测性能进行了分析和仿真,结果表明在复杂信号环境中当常规检测器由于虚警概率很高失效时,该检测器可以同时获得较低的虚警概率和较高的检测概率,因而适用于复杂信号环境中非合作突发信号检测。该检测器运算量小,易于实时实现。展开更多
文摘针对浅层学习模型对风速预测存在较大误差的缺陷,提出一种基于小波变换和深度信念网络(wavelet based deep belief netw ork,WDBN)的风速预测模型。首先利用小波变换将原始风速序列分解成不同的频率序列;然后针对各频率序列,根据逐层训练法则设计深度信念网络模型;最后采用小波反变换对不同频率序列重构,得出最终的风速预测结果。选用某风电场2013年1月和7月的数据对WDBN模型的风速预测进行仿真分析,并与自回归滑动平均法、反向传播神经网络法、Morlet小波神经网络法的预测结果进行对比,结果表明WDBN模型可以更好地学习风速所具有的非线性和非平稳特征,具有较高的预测精度。
基金Supported partially by Chinese NNSF grants 19731010the knowledge innovation program of CAS.
文摘In a composite-step approach, a step Sk is computed as the sum of two components Uk and hk. The normal component Vk, which is called the vertical step, aims to improve the linearized feasibility, while the tangential component hk, which is also called horizontal step, concentrates on reducing a model of the merit functions. As a filter method, it reduces both the infeasibility and the objective function. This is the same property of these two methods. In this paper, one concerns the composite-step like filter approach. That is, a step is tangential component hk if the infeasibility is reduced. Or else, Sk is a composite step composed of normal component Uk, and tangential component hk.
文摘In this paper, we account for this subject: how to de sign a pattern, it can track the state of the equipment of some organizations su ch as enterprise, organ, laboratory, school etc. We present an analysis pattern, which describes the whole procedure of managing the equipment and record the us ing of the equipment. It not only can track the quantity and location of the equ ipment of the whole organization, the more important is it can update the state of the equipment at real-time automatically. First,we design the static diagram(i.e. using UML class diagram to describe the basic state of the equipment). Then we consider its dynamic aspect, i.e., how th e state of the equipment to get changed according to the time. We use UML sequen ce diagram and state diagram to respectively describe the procedure of after pur chasing, transferring and discarding as useless of the equipment. Obviously, the manager can update the quantity and location of the equipment automatic ally. We character this pattern from the following five aspects: Problem: How the enterprise, organ, laboratory and school to track the quantity and location of the equipment. Circumstance: In some organizations, especially, in the manufactory or laborator y, when the number of the quantity and type is large or the distribution of the equipment is dispersed, they want to be able to track the quantity and location of the equipment. Forces: First, it is possible to the equipment be transferred or be discarded, n o matter when and where, the organization must be able to track the factual quan tity and location.Second, the solution must describe a basic semantic unit, that is, the solution must simple enough to apply it to various of circumstance, whi ch is the base of reusing. Third, the solution must include the interpret of the factual document. Solution: This part, we start with the basic demands, first using UML class diag ram to describe the basic pattern, which is an atomic pattern. Then using UML se quence diagram and state diagram to respectively de
文摘现有未知突发信号检测算法是基于噪声加单一突发信号的简单假设的,在实际复杂信号环境会产生大量虚警而失效。针对实际非合作突发通信信号的检测环境除噪声外还包含多个连续信号和一些短突发干扰信号,建立了复杂信号环境模型,提出了适用于此环境的基于短时傅里叶变换(short time Fourier transform,STFT)的时序检测器。该检测器利用突发通信信号时间上短持续的特点剔除连续信号和短突发干扰造成的虚警。对该检测器的检测性能进行了分析和仿真,结果表明在复杂信号环境中当常规检测器由于虚警概率很高失效时,该检测器可以同时获得较低的虚警概率和较高的检测概率,因而适用于复杂信号环境中非合作突发信号检测。该检测器运算量小,易于实时实现。
