linear adj. 1.線的,直線的。 2.長度的。 3.【數(shù)學(xué)】一次的,線性的。 4.【動(dòng)、植】線狀的;細(xì)長的。 5.由線條組成的,以線條為主的,強(qiáng)調(diào)線條的。 linear amplification 直線放大。 a linear equation 一次方程式。 a linear leaf 線形葉。 linear arts 線條藝術(shù)。
After analyzing the difficulties that lies in comprehensive evaluation of the anti-jamming ability of radar net, the paper discusses the probability of solving the problem using the excellent non-linear function approaching speciality owned by bp neural network, and analyses various problem that required to pay attention to during the modeling and evaluating process, emphasizes particularly on determining the input index aggregate together with output index aggregate and training arithmetic and validation when constructing the evaluation model, the potential use of established model is also probed into 摘要在分析目前綜合評(píng)估雷達(dá)網(wǎng)抗干擾性能時(shí)遇到困難后,探討了利用bp人工神經(jīng)網(wǎng)絡(luò)良好的非線性函數(shù)逼近特性來解決這一問題的可行性,并分析了在用這種方法建模評(píng)估的過程中需要注意的各種問題;其中側(cè)重對在具體構(gòu)造評(píng)估模型時(shí)輸入輸出評(píng)估指標(biāo)集的確定方法和模型的訓(xùn)練驗(yàn)證方法進(jìn)行了研究,并進(jìn)一步分析了建立的模型可能的用途。
The approximation property of projection pursuit wavelet neural network ( ppwnn ) which is applied to non-linear function is studied, the convergence rate is given in this paper also . 3 . we demonstrate projection pursuit wavelet neural network ( wppnn ) has a good applicability by the approximation of five non-linear functions and the prediction of sunport and chaos time series and the use of edge detection 主要工作如下:(1)建立了投影尋蹤小波神經(jīng)網(wǎng)絡(luò)的數(shù)學(xué)模型、拓?fù)浣Y(jié)構(gòu)及非線性學(xué)習(xí)機(jī)理;(2)證明了投影尋蹤小波神經(jīng)網(wǎng)絡(luò)可以逼近非線性函數(shù),并給出了其收斂速度;(3)通過投影尋蹤小波神經(jīng)絡(luò)對五種非線性函數(shù)的逼近和對太陽黑子、混沌時(shí)間序列的模擬預(yù)報(bào)以及投影尋蹤小波學(xué)習(xí)網(wǎng)絡(luò)在圖象邊緣檢測中的應(yīng)用,說明該網(wǎng)絡(luò)具有很強(qiáng)的實(shí)用性。
The approximation property of projection pursuit wavelet neural network ( ppwnn ) which is applied to non-linear function is studied, the convergence rate is given in this paper also . 3 . we demonstrate projection pursuit wavelet neural network ( wppnn ) has a good applicability by the approximation of five non-linear functions and the prediction of sunport and chaos time series and the use of edge detection 主要工作如下:(1)建立了投影尋蹤小波神經(jīng)網(wǎng)絡(luò)的數(shù)學(xué)模型、拓?fù)浣Y(jié)構(gòu)及非線性學(xué)習(xí)機(jī)理;(2)證明了投影尋蹤小波神經(jīng)網(wǎng)絡(luò)可以逼近非線性函數(shù),并給出了其收斂速度;(3)通過投影尋蹤小波神經(jīng)絡(luò)對五種非線性函數(shù)的逼近和對太陽黑子、混沌時(shí)間序列的模擬預(yù)報(bào)以及投影尋蹤小波學(xué)習(xí)網(wǎng)絡(luò)在圖象邊緣檢測中的應(yīng)用,說明該網(wǎng)絡(luò)具有很強(qiáng)的實(shí)用性。
In the research of the algorithms and theory of temporal difference learning, a new class of multi-step learning prediction algorithms based on linear function approximators and recursive least squares methods is proposed, which are called the rls-td ( t ) learning algorithm . the convergence with probability one of the rls-td ( t ) algorithm is proved for ergodic markov chains, and the conditions for convergence are analyzed 在時(shí)域差值學(xué)習(xí)(temporaldifferencelearning)學(xué)習(xí)算法和理論方面,首次提出了一種基于線性值函數(shù)逼近的多步遞推最小二乘td()(rls-td())學(xué)習(xí)算法,并分析和證明了該算法在求解遍歷markov鏈學(xué)習(xí)預(yù)測問題中的收斂條件和一致收斂性。
Thirdly, introduce the supplier classification system and its evaluation factors using the ahp and the linear function to get the distribution situation in a two-dimension matrix model . at last, this thesis gives a demonstration in the xizi-otis co . ltd, and compare with their model in use with the hope to improve the model and can make some sense to companies 最后,通過對杭州西子otis電梯有限公司供應(yīng)商分類管理的實(shí)證分析,對該方法進(jìn)行了檢驗(yàn),并與西子otis的供應(yīng)商管理現(xiàn)狀相比較,通過分析驗(yàn)證結(jié)果與實(shí)際情況的差異,提出了進(jìn)一步改善本文研究方法的想法。
The reasons focusing only on ray travel times are " noise tolerance could be enhanced and the sensitivity to the environment model mismatch may be reduced by neglecting amplitude information; phase ambiguity is eliminated by employing the group delay of arriving wave packets; travel times are pseudo-linear functions of sound speed profiles of the medium, and the perturbation of ray travel time is much smaller than sound pressure's 之所以選擇聲傳播時(shí)間的原因是:噪聲容限可以得到提高,并且傳播時(shí)間的計(jì)算可以不考慮聲壓的信息,從而降低了對環(huán)境模型失配的敏感性;相位的起伏也可通過到達(dá)包絡(luò)的群延時(shí)來消除;傳播時(shí)間是介質(zhì)聲速分布的偽線性函數(shù)并且傳播時(shí)間的起伏遠(yuǎn)小于聲壓的起伏。
We consider the traveling waves of a delay cnn system with the output function f given as a non-linear function in the latter . we show the existence of monotone traveling waves solutions and describe the global structure of traveling waves such as monotone, damped oscillation, periodic oscillation and unbounded etc as the wave speed c is classified in r 第二部分我們考慮具時(shí)滯非線性cnn系統(tǒng)模型,得出方程單調(diào)行波解存在,并將波速cr分成若干個(gè)區(qū)間,得出c在各個(gè)區(qū)間時(shí)行波解曲線相應(yīng)的類型包括單調(diào)、衰減振動(dòng)、周期振動(dòng)、無界等類型。
(2 ) runoff-yielding time on grassland with a higher cover ratio primarily depended on initial water content, and their relationship might be described by a power function too . surface crust could easily take place on bare land, so its runoff-yielding time primarily depended on rainfall intensity, and their relationship might be described by a linear function (2)植被覆蓋度較高的荒草地,產(chǎn)流時(shí)間主要取決于初始含水量,兩者的關(guān)系可用冪函數(shù)描述;裸地因降雨產(chǎn)生地表結(jié)皮,產(chǎn)流時(shí)間主要取決于降雨強(qiáng)度,兩者的關(guān)系可用線性描述。
Details are as follows : we deal with properties of bilevel linear programming and prove the equivalence of bilevel linear programming and optimization over the efficient set . a class of multi-objective tow level programming, i . e . the upper-level is single objective and the lower-level is linear multi-objective, is mainly discussed . it can be converted into the optimization over the efficient set with parameter and an algorithm is given with its finite termination being proved; when the upper-level is linear function, an exact penalty function algorithm is given 分層(分級(jí))遞階系統(tǒng)是社會(huì)組織管理的主要形式,多層規(guī)劃是研究這類系統(tǒng)優(yōu)化問題的基本模型,其鮮明的實(shí)際背景和廣泛的應(yīng)用前景引起了人們的廣泛關(guān)注,成為一個(gè)新興的活躍的研究領(lǐng)域,本論文研究了二層規(guī)劃中的若干問題,主要工作如下:討論了二層線性規(guī)劃的性質(zhì),并證明了它與零有效集上優(yōu)化問題的等價(jià)性;對一類二層多目標(biāo)規(guī)劃(上層為單目標(biāo)規(guī)劃、下層為線性多目標(biāo)規(guī)劃的問題)進(jìn)行了探討,將其轉(zhuǎn)化為含參變量的有效集上的優(yōu)化問題,進(jìn)而給出了一種算法,并證明了該算法的有限終止性;當(dāng)上層為線性單目標(biāo)時(shí),給出了一種罰函數(shù)方法
Finally, take example for a non-linear function, method mentioned in this paper is used to design wavelet neural network to approximate this function . the computer simulations confirm the method that is brought out in this paper is useful, and prove that wavelet neural network has not only fast convergence and better precision of approximation, but also good capability of forecasting and escaping error 最后,對于一個(gè)實(shí)際的非線性函數(shù),用本文介紹的方法來設(shè)計(jì)小波神經(jīng)網(wǎng)絡(luò)來逼近函數(shù),仿真結(jié)果表明該方法的有效性,并且表明小波神經(jīng)網(wǎng)絡(luò)在函數(shù)逼近上,網(wǎng)絡(luò)的收斂速度快,逼近精度高的特點(diǎn),并且網(wǎng)絡(luò)具有很好的泛化能力和容錯(cuò)性。