国产日韩欧美一区二区三区三州_亚洲少妇熟女av_久久久久亚洲av国产精品_波多野结衣网站一区二区_亚洲欧美色片在线91_国产亚洲精品精品国产优播av_日本一区二区三区波多野结衣 _久久国产av不卡

?

基于PSO-RBF神經(jīng)網(wǎng)絡(luò)的海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)預(yù)測(cè)

2019-02-19 02:29楊潔程曉健穆彥斌
現(xiàn)代電子技術(shù) 2019年3期
關(guān)鍵詞:粒子群算法遺傳算法神經(jīng)網(wǎng)絡(luò)

楊潔 程曉健 穆彥斌

關(guān)鍵詞: 海戰(zhàn)場(chǎng); 電磁態(tài)勢(shì); 神經(jīng)網(wǎng)絡(luò); 粒子群算法; 模擬退火法; 遺傳算法

中圖分類號(hào): TN911.1?34; TP311.54 ? ? ? ? ? ? ? ? ? 文獻(xiàn)標(biāo)識(shí)碼: A ? ? ? ? ? ? ? ?文章編號(hào): 1004?373X(2019)03?0001?05

Abstract: A sea battlefield electromagnetic state prediction method based on improved particle swarm optimization (PSO)algorithm optimizing radial basis function (RBF) neural network is proposed to solve the prediction problem of sea battlefield electromagnetic state. The adaptive inertia weight, simulated annealing method and genetic algorithm are used in the method to improve the conventional PSO algorithm, and its search accuracy and speed. The improved PSO algorithm is used to optimize the parameters of RBF neural network, which can improve the learning efficiency and prediction accuracy of the network. The simulation prediction is carried out for the non?linear mapping relationship between the electromagnetic state values of the sea battlefield. The experimental results show that the method can improve the prediction accuracy of the sea battlefield electromagnetic state effectively, and has strong applicability.

Keywords: sea battlefield; electromagnetic state; neural network; particle swarm optimization algorithm; simulated annealing method; genetic algorithm

0 ?引 ?言

海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)感知是一種通過(guò)對(duì)海戰(zhàn)場(chǎng)電磁環(huán)境要素的獲取、理解、預(yù)測(cè)而形成易于指揮員準(zhǔn)確認(rèn)識(shí)海戰(zhàn)場(chǎng)電磁環(huán)境并能輔助其決策的方法[1]?,F(xiàn)有的態(tài)勢(shì)評(píng)估方法大多只能提供給指揮員過(guò)去和當(dāng)前的海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)情況,無(wú)法預(yù)測(cè)下一階段態(tài)勢(shì)變化情況,使得己方在未來(lái)戰(zhàn)爭(zhēng)中處于被動(dòng)狀態(tài)。因此,海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)預(yù)測(cè)成為未來(lái)戰(zhàn)場(chǎng)中亟待解決的問(wèn)題。

目前國(guó)內(nèi)外對(duì)于海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)的研究主要集中在電磁環(huán)境可視化[2]、電磁環(huán)境復(fù)雜度評(píng)估[3]、輻射源識(shí)別[4]等方面,缺乏生成系統(tǒng)海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)的技術(shù)手段。文獻(xiàn)[1]提出了海戰(zhàn)場(chǎng)電磁感知的基本模型,但并未對(duì)態(tài)勢(shì)理解域中的態(tài)勢(shì)預(yù)測(cè)作進(jìn)一步分析。文獻(xiàn)[5]將博弈論應(yīng)用于戰(zhàn)場(chǎng)通信對(duì)抗態(tài)勢(shì)預(yù)測(cè)中,但預(yù)測(cè)結(jié)果誤差較大。徑向基函數(shù)(Radial Basis Function,RBF)神經(jīng)網(wǎng)絡(luò)具有收斂速度快、結(jié)構(gòu)簡(jiǎn)單、非線性映射能力好等特點(diǎn)[6],已廣泛應(yīng)用于模式識(shí)別[7]、網(wǎng)絡(luò)安全態(tài)勢(shì)預(yù)測(cè)[8]等領(lǐng)域。同時(shí),為了提高RBF神經(jīng)網(wǎng)絡(luò)性能,國(guó)內(nèi)學(xué)者利用粒子群算法(Particle Swarm Optimization,PSO)的搜索能力和RBF神經(jīng)網(wǎng)絡(luò)的非線性映射能力,提出改進(jìn)粒子群算法優(yōu)化RBF神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)模型[9]。

為了準(zhǔn)確把握海戰(zhàn)場(chǎng)電磁發(fā)展態(tài)勢(shì),在已有研究成果的基礎(chǔ)上,提出一種基于改進(jìn)PSO算法優(yōu)化RBF神經(jīng)網(wǎng)絡(luò)的海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)預(yù)測(cè)方法。該方法首先對(duì)海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)要素進(jìn)行分析,繼而獲得海戰(zhàn)場(chǎng)電磁整體態(tài)勢(shì)值,然后采用改進(jìn)的粒子群算法優(yōu)化RBF神經(jīng)網(wǎng)絡(luò),尋找海戰(zhàn)場(chǎng)電磁值之間的非線性映射關(guān)系,對(duì)未來(lái)時(shí)刻海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)進(jìn)行預(yù)測(cè)。

1 ?海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)預(yù)測(cè)框架

海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)值是在對(duì)海戰(zhàn)場(chǎng)電磁環(huán)境物理特性和電磁環(huán)境中電子設(shè)備用頻效能分析的基礎(chǔ)上,通過(guò)一定的數(shù)學(xué)模型進(jìn)行計(jì)算,將人們不易理解的海戰(zhàn)場(chǎng)電磁環(huán)境和戰(zhàn)場(chǎng)態(tài)勢(shì)信息歸并融合成人們?nèi)菀桌斫夂徒邮艿臄?shù)值。這些數(shù)值能夠客觀實(shí)時(shí)反映海戰(zhàn)場(chǎng)電磁域中戰(zhàn)場(chǎng)態(tài)勢(shì)情況,其大小取決于海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)要素。針對(duì)電磁環(huán)境特點(diǎn),將電磁態(tài)勢(shì)劃分為一般態(tài)勢(shì)和相對(duì)態(tài)勢(shì)兩部分。一般態(tài)勢(shì)如海戰(zhàn)場(chǎng)區(qū)域內(nèi)電磁信號(hào)的空間覆蓋率、時(shí)間占用率、頻段占用率、平均功率密度譜等[10];相對(duì)態(tài)勢(shì)如探測(cè)雷達(dá)的發(fā)現(xiàn)目標(biāo)概率和最大探測(cè)距離,制導(dǎo)系統(tǒng)對(duì)目標(biāo)的跟蹤精度和制導(dǎo)概率,通信系統(tǒng)之間的誤信率、誤碼率以及電子設(shè)備和系統(tǒng)在電子干擾和反輻射攻擊中的生存能力等[11]。電磁態(tài)勢(shì)評(píng)估指標(biāo)體系如圖1所示。

同時(shí)為了驗(yàn)證所提算法的優(yōu)越性,采用IMPSO?RBF預(yù)測(cè)模型[9]及SACPSO?RBF預(yù)測(cè)模型[12]進(jìn)行相同的實(shí)驗(yàn)。網(wǎng)絡(luò)訓(xùn)練過(guò)程中最優(yōu)適應(yīng)度值曲線如圖3所示,預(yù)測(cè)結(jié)果如圖4所示。

由圖3所示,盡管三種預(yù)測(cè)模型在網(wǎng)絡(luò)訓(xùn)練過(guò)程中最佳適應(yīng)度值都可以很快收斂到最小值,但相對(duì)于其他兩種預(yù)測(cè)模型,該預(yù)測(cè)模型可以更快找到態(tài)勢(shì)值之間的非線性映射關(guān)系。其原因在于本文方法能夠根據(jù)粒子群中粒子的適應(yīng)度值自適應(yīng)賦予其移動(dòng)速度權(quán)重,能更快地尋找到最佳粒子位置,因此,加快了優(yōu)化后的RBF網(wǎng)絡(luò)預(yù)測(cè)模型的收斂速度。

由圖4中的預(yù)測(cè)曲線可以看出,三種預(yù)測(cè)模型都取得了一定的預(yù)測(cè)效果。如圖5所示,本文方法的預(yù)測(cè)效果更好,更符合真實(shí)電磁態(tài)勢(shì)變化趨勢(shì)。這是因?yàn)楸疚姆椒ú捎媚M退火法避免了粒子群算法在搜索過(guò)程中陷入局部極小值的問(wèn)題,并采用遺傳算法中的交叉、變異操作提高了種群多樣性,提高了PSO算法在全局最優(yōu)解的搜索能力。

為了進(jìn)一步體現(xiàn)本文方法的優(yōu)越性,分別計(jì)算了三種預(yù)測(cè)模型的均方根誤差(RMSE)、平均相對(duì)誤差(MAPE),如表1所示。

從表1中可以看出,相對(duì)于IMPSO?RBF預(yù)測(cè)模型及SACPSO?RBF預(yù)測(cè)模型,本文預(yù)測(cè)模型得到的電磁態(tài)勢(shì)值的均方根誤差(RMSE),平均相對(duì)誤差(MAPE)均明顯降低。

4 ?結(jié) ?語(yǔ)

針對(duì)海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)的預(yù)測(cè)問(wèn)題,本文提出一種基于改進(jìn)PSO優(yōu)化RBF神經(jīng)網(wǎng)絡(luò)的海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)預(yù)測(cè)方法。通過(guò)對(duì)海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)要素的分析,生成能客觀反映海戰(zhàn)場(chǎng)電磁域的整體態(tài)勢(shì)值,利用改進(jìn)PSO算法優(yōu)化RBF神經(jīng)網(wǎng)絡(luò)參數(shù),并與其他預(yù)測(cè)模型的測(cè)試結(jié)果對(duì)比,該方法可以取得更高的預(yù)測(cè)精度,在海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)預(yù)測(cè)領(lǐng)域具有一定的應(yīng)用價(jià)值。

參考文獻(xiàn)

[1] 周倜,王小非,陳煒.海戰(zhàn)場(chǎng)電磁態(tài)勢(shì)感知模型[J].火力與指揮控制,2013,38(8):1?5.

ZHOU Ti, WANG Xiaofei, CHEN Wei. Electromagnetic situation perception model for sea battlefield [J]. Fire control & command control, 2013, 38(8): 1?5.

[2] TANG D, HAN H, YUAN K. Research on the essence and visualization description method of battlefield electromagnetic environment [J]. Ordnance industry automation, 2014(11): 57?59.

[3] WANG F, HAN H, WANG J, et al. The complexity evaluation method of electromagnetic environment based on statistical characteristics analysis [J]. Applied mechanics & materials, 2013, 321/324: 779?784.

[4] 陳求,戎華,譚亮亮.基于最小二乘法的雷達(dá)輻射源精確識(shí)別指標(biāo)權(quán)重確定方法[J].艦船電子對(duì)抗,2016,39(4):52?55.

CHEN Qiu, RONG Hua, TAN Liangliang. A method for determining precise identification indexes of radar emitters based on least squares [J]. Shipboard electronic countermeasure, 2016, 39(4): 52?55.

[5] 馮德俊,朱江,李方偉.戰(zhàn)場(chǎng)電磁態(tài)勢(shì)感知關(guān)鍵技術(shù)研究[J].數(shù)字通信,2013,40(5):20?23.

FENG Dejun, ZHU Jiang, LI Fangwei. Research on key technologies of electromagnetic situation sensing in battlefield [J]. Digital communication, 2013, 40(5): 20?23.

[6] HAN H G, QIAO J F. Prediction of activated sludge bulking based on a self?organizing RBF neural network [J]. Journal of process control, 2012, 22(6): 1103?1112.

[7] 杜剛,何朔,于海鵬.基于徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)的空間碎片撞擊模式識(shí)別研究[J].航天器環(huán)境工程,2015,32(4):357?360.

DU Gang, HE Shuo, YU Haipeng. Research on pattern recognition of space debris impact based on radial basis function neural network [J]. Spacecraft environment engineering, 2015, 32(4): 357?360.

[8] 李方偉,鄭波,朱江,等.一種基于AC?RBF神經(jīng)網(wǎng)絡(luò)的網(wǎng)絡(luò)安全態(tài)勢(shì)預(yù)測(cè)方法[J].重慶郵電大學(xué)學(xué)報(bào)(自然科學(xué)版),2014,26(5):576?581.

LI Fangwei, ZHENG Bo, ZHU Jiang, et al. An approach to forecast the network security situation based on AC?RBF neural network [J]. Journal of Chongqing University of Posts and Telecommunications (natural science edition), 2014, 26(5): 576?581.

[9] 夏軒,許偉明.改進(jìn)的粒子群算法對(duì)RBF神經(jīng)網(wǎng)絡(luò)的優(yōu)化[J].計(jì)算機(jī)工程與應(yīng)用,2012,48(5):37?40.

XIA Xuan, XU Weiming. Optimization of RBF neural network based on improved particle swarm optimization [J]. Journal of computer engineering and applications, 2012, 48(5): 37?40.

[10] CAI X F, SONG J S. Analysis of complexity in battlefield electromagnetic environment [C]// 2009 IEEE Conference on Industrial Electronics and Applications. Xian, China: IEEE, 2009: 2440?2442.

[11] 高波,馬向玲,隋江波.海戰(zhàn)復(fù)雜電磁環(huán)境分析[J].火力與指揮控制,2013,38(3):1?4.

GAO Bo, MA Xiangling, SUI Jiangbo. Analysis of complex electromagnetic environment in sea battle [J]. Fire control & command control, 2013, 38(3): 1?4.

[12] 張義,田愛(ài)奎,韓士元.一種自適應(yīng)的混沌粒子群優(yōu)化RBF神經(jīng)網(wǎng)絡(luò)算法[J].重慶理工大學(xué)學(xué)報(bào),2015,29(11):126?130.

ZHANG Yi, TIAN Aikui, HAN Shiyuan. An adaptive RBF neural network algorithm based on chaotic particle swarm optimization [J]. Journal of Chongqing Institute of Technology, 2015, 29(11): 126?130.

[13] CHEN G C, YU J S. Particle swarm optimization algorithm [J]. Information & control, 2005, 306: 1369?1372.

[14] 董俊,洪麗娜,汪連棟,等.多輻射平臺(tái)運(yùn)動(dòng)區(qū)域電磁環(huán)境預(yù)測(cè)方法[J].現(xiàn)代防御技術(shù),2016,44(2):190?196.

DONG Jun, HONG Lina, WANG Liandong, et al. An electromagnetic environment prediction method in multi?radiation platform motion area [J]. Modern defence technology, 2016, 44(2): 190?196.

猜你喜歡
粒子群算法遺傳算法神經(jīng)網(wǎng)絡(luò)
神經(jīng)網(wǎng)絡(luò)抑制無(wú)線通信干擾探究
基于自適應(yīng)遺傳算法的CSAMT一維反演
一種基于遺傳算法的聚類分析方法在DNA序列比較中的應(yīng)用
基于遺傳算法和LS-SVM的財(cái)務(wù)危機(jī)預(yù)測(cè)
電力市場(chǎng)交易背景下水電站優(yōu)化調(diào)度研究
基于粒子群算法的產(chǎn)業(yè)技術(shù)創(chuàng)新生態(tài)系統(tǒng)運(yùn)行穩(wěn)定性組合評(píng)價(jià)研究
交通堵塞擾動(dòng)下多車場(chǎng)車輛路徑優(yōu)化
基于神經(jīng)網(wǎng)絡(luò)的拉矯機(jī)控制模型建立
基于改進(jìn)的遺傳算法的模糊聚類算法
復(fù)數(shù)神經(jīng)網(wǎng)絡(luò)在基于WiFi的室內(nèi)LBS應(yīng)用