(11)
5案例分析
通過對案例的發(fā)展規(guī)劃過程進(jìn)行研究,可分析所建模型的有效性。假定當(dāng)前需要規(guī)劃發(fā)展某偵察預(yù)警監(jiān)視體系,該體系的能力需求內(nèi)容、可選裝備類型、裝備與能力之間的映射關(guān)系如圖5所示。
(1) 參數(shù)設(shè)定
由圖5可知,N=7,M=7。假設(shè)采集了近3年的ω取值,見表1。
表1 近3年ω參數(shù)設(shè)定
則υ=(0.21,0.15,0.13,0.09,0.08,0.16,0.18)。令
圖5 決策變量-能力需求之間的關(guān)系Fig.5 Relationship between decision variable and capability requirement
表和參數(shù)設(shè)定
(2) 算法驗證
在這里采用改進(jìn)的遺傳算法對模型進(jìn)行驗證。在文獻(xiàn)[13]中,基于捕食搜索策略(predatory search,PS)對遺傳算法(genetic algorithm,GA)進(jìn)行了改進(jìn),具體思想為:將捕食搜索策略應(yīng)用到遺傳算法中,克服遺傳算法局部搜索能力弱的問題。首先以較大的交叉概率和較小的變異概率進(jìn)行全局搜索;一旦發(fā)現(xiàn)一個較優(yōu)解,則改為以較小的交叉概率和較大的變異概率進(jìn)行局部搜索; 如果在一定次數(shù)
的搜索過程中較優(yōu)解得不到改善,則恢復(fù)以較大的交叉概率和較小的變異概率進(jìn)行全局搜索?;诓妒乘阉鞑呗缘倪z傳算法(PSGA)過程如圖6所示。
圖6 PSGA算法流程圖Fig.6 Flow diagram of PSGA arithmetic
為了便于比較,這里采用PSGA算法與傳統(tǒng)遺傳算法(GA)、蟻群算法(ant colony algorithm, CA)、微分進(jìn)化算法(differential evolution algorithm,DE),分別對這一案例進(jìn)行求解。根據(jù)文獻(xiàn)[14-16],各算法的參數(shù)設(shè)置見表3。
表3 PSGA/GA/CA/DE參數(shù)設(shè)置
首先分析算法的求解穩(wěn)定性。分別對4種算法獨立運(yùn)行10次,獲得每次運(yùn)算的體系總軍事價值最優(yōu)值,見圖7。對10次運(yùn)算所得最優(yōu)值的最大值、均值和方差進(jìn)行計算,結(jié)果見表4??梢姡琍SGA的方差值是4類算法中最小的,代表PSGA求解穩(wěn)定性最好。
表4 PSGA/GA/CA/DE獨立運(yùn)算結(jié)果
圖7 PSGA,GA,CA,DE求解穩(wěn)定性比較Fig.7 Stability comparison among PSGA,GA, CA, and DE
然后分析算法的求解效率性。將4類算法10次獨立運(yùn)行中得出最大體系總軍事價值的求解過程
提取出來進(jìn)行比較,見圖8??梢?,在收斂速度方面,PSGA在第64次達(dá)到收斂,而GA,CA,DE 3類算法分別在81,90,72次才達(dá)到收斂。而在最優(yōu)值的搜尋上,PSGA,GA,CA,DE 4類算法分別是0.710 0,0.521 0,0.532 6,0.624 1。所以,無論是收斂速度,還是最優(yōu)值的搜尋能力,PSGA均優(yōu)于其他3類算法,PSGA具有更高的求解效率性。
圖8 PSGA,GA,CA,DE求解效率性比較Fig.8 Efficiency comparison among PSGA, GA, CA, and DE
(3) 規(guī)劃方案分析
分析4類算法所得的最優(yōu)發(fā)展規(guī)劃方案,見表5??芍?類算法都傾向于優(yōu)先發(fā)展預(yù)警機(jī)、偵察衛(wèi)星2類武器裝備。并且PSGA算法在預(yù)警機(jī)裝備上投入資金最多,所得體系總軍事價值也最大。規(guī)劃方案結(jié)果的得出與前面設(shè)置的參數(shù)有關(guān),實際操作時應(yīng)根據(jù)作戰(zhàn)實際嚴(yán)格設(shè)置參數(shù),才能得出有效的結(jié)論。
表5 最優(yōu)發(fā)展規(guī)劃方案
6結(jié)束語
將Markowitz理論應(yīng)用于武器裝備體系規(guī)劃,既為Markowitz投資組合理論開拓一個新的應(yīng)用領(lǐng)域,也為武器裝備體系發(fā)展規(guī)劃找到一個新的定量決策模型,且較好地解決了戰(zhàn)略走勢變化帶來的不確定影響這一問題,所建模型可為類似問題規(guī)劃提供模型參考。
本文從理論和方法的角度對武器裝備體系規(guī)劃做了探索。下一步的研究應(yīng)包括:如何用數(shù)學(xué)方法科學(xué)精確地描述軍事價值;如何建立可操作性強(qiáng)、更反映規(guī)劃實際的數(shù)學(xué)模型;如何確定投入資金與武器能力值之間映射關(guān)系的函數(shù)——即文中的Logistic函數(shù)中的變量系數(shù);如何綜合考慮并描述形勢變化帶來的其他不確定影響因素;如何構(gòu)建包含更多規(guī)劃目標(biāo)和約束條件的模型;如何改進(jìn)算法以實現(xiàn)更快更穩(wěn)的尋優(yōu)。
參考文獻(xiàn):
[1]李仁傳,張合勇,殷燕. 對武器裝備體系結(jié)構(gòu)優(yōu)化幾個基本問題的理性認(rèn)識[J]. 軍事運(yùn)籌與系統(tǒng)工程,2011,25(2):5-10.
LI Ren-chuan,ZHANG He-yong,YIN Yan. An Understanding on Several Basic Problems of Weapon System of System Structural Optimization [J]. Military Operations Research and Systems Engineering,2011,25(2):5-10.
[2]周宇,譚躍進(jìn),姜江,等. 面向能力需求的武器裝備體系組合規(guī)劃模型與算法[J]. 系統(tǒng)工程理論與實踐,2013,33(3):809-816.
ZHOU Yu,TAN Yue-jin,JIANG Jiang,et al. Capability Requirements Oriented Weapon System of Systems Portfolio Planning Model and Algorithm [J]. Systems Engineering—Theory and Practice,2013,33(3):809-816.
[3]熊健,趙青松,葛冰峰,等. 基于多目標(biāo)優(yōu)化模型的武器裝備體系能力規(guī)劃[J]. 國防科技大學(xué)學(xué)報,2011,33(3):140-144.
XIONG Jian,ZHAO Qing-song,GE Bing-feng,et al. Weapon Equipment System-of-Systems Capability Planning Based on Multi-Objective Optimization Model [J]. Journal of National University of Defense Technology,2011,33(3):140-144.
[4]LEE J,KANG S H,ROSENBERGER J,et al. A Hybrid Approach of Goal Programming for Weapon Systems Selection[J]. Computers and Industrial Engineering,2010,58(3):521-527.
[5]SNYDER D,MILLS P,RESNICK A C,et a1. Assessing Capabilities and Risks in Air Force Programming[R]. Pittsburgh:RAND Corporation,2009.
[6]STRUBE C M,LOREN J R.Portfolio Influences on Air Force Capabilities-Based Assessment and Capabilities-Based Planning Activities[C]∥Proceedings of the 6th International Conference on System of Systems Engineering,Albuquerque,New Mexico,USA,2011:83-89.
[7]United States Government Accountability Omce. Defense Acquisitions Assessments of Selected Weapon Programs[EB/OL].[2012-4-18](2014-11-28).http://
www.gao.gov/assets/320/317081.pdf.
[8]GREINER M A,F(xiàn)OWLER J W,SHUNK D L,et al.A hybrid Approach Using the Analytic Hierarchy Process and Integer Programming to Screen Weapon Systems Projects[J].IEEE Transactions on Engineering Management,2003,50(2):192-203.
[9]張媛,劉文彪,張立民. 基于主客觀綜合賦權(quán)的CGF態(tài)勢評估建模研究[J]. 系統(tǒng)工程與電子技術(shù),2013,35(1):85-90.
ZHANG Yuan,LIU Wen-biao,ZHANG Li-min. Situation Assessment Modeling for CGF Base on the Subjective and Objective Integrated Weight [J]. Systems Engineering and Electronics,2013,35(1):85-90.
[10]劉旭,李為民,吳曉東. 雷達(dá)抗干擾性能多層次灰色評估模型研究[J]. 現(xiàn)代防御技術(shù),2011,39(6):179-184.
LIU Xu,LI Wei-min,WU Xiao-dong. Multilevel Gray Evaluation Model for Radar Antijamming Capability [J]. Modern Defence Technology,2011,39(6):179-184.
[11]吳曉. Markowitz投資組合理論在房地產(chǎn)投資組合決策中的應(yīng)用[J]. 深圳職業(yè)技術(shù)學(xué)院學(xué)報,2011(1):36-42.
WU Xiao. Application of Markowitz Portfolio Theory in Real Estate Portfolio Decision [J]. Journal of Shenzhen Polytechnic,2011(1):36-42.
[12]楊明輝,張智光,任百林,等. Markowitz組合證券投資決策模型的修正[J]. 南京林業(yè)大學(xué)學(xué)報:自然科學(xué)版,2005,29(1):51-54.
YANG Ming-hui, ZHANG Zhi-guang, REN Bai-lin, et al. A Study on Revising Markowitz’s Portfolio Selection Model [J]. Journal of Nanjing Forestry University:Natural Sciences ed, 2005,29(1):51-54.
[13]劉旭,李為民,宋文靜. 考慮發(fā)射區(qū)部分重疊的防空作戰(zhàn)目標(biāo)分配[J]. 空軍工程大學(xué)學(xué)報:自然科學(xué)版,2013,14(6):30-34.
LIU Xu,LI Wei-min,SONG Wen-jing. Target Assignment of Defense Combat in Overlapping Shooting-Area [J]. Journal of Air Force Engineering University:Natural Science ed,2013,14(6):30-34.
[14]余家祥,趙曉哲,史紅權(quán),等. 基于遺傳算法的編隊區(qū)域防空武器分配方法[J]. 現(xiàn)代防御技術(shù),2013,41(1):82-87.
YU Jia-xiang,ZHAO Xiao-zhe,SHI Hong-quan,et al. Air Defense Weapon Allocation Method for Warship Task Formation Based on Genetic Algorithm [J]. Modern Defence Technology,2013,41(1):82-87.
[15]李紅亮,宋貴寶,曹延杰. 多反艦導(dǎo)彈攻擊多目標(biāo)協(xié)同航路規(guī)劃[J]. 系統(tǒng)工程與電子技術(shù),2013,35(10):2013-2019.
LI Hong-liang,SONG Gui-bao,CAO Yan-jie. Cooperative Path Planning of Multiple Anti-Ship Missiles to Multiple Targets [J]. Systems Engineering and Electronics,2013,35(10):2013-2019.
[16]黃仁全,李為民,周曉光,等. 基于微分進(jìn)化算法的防空導(dǎo)彈火力分配[J]. 空軍工程大學(xué)學(xué)報:自然科學(xué)版,2009,10(5):41-44.
HUANG Ren-quan,LI Wei-min,ZHOU Xiao-guang,et al. Research on Firepower Distribution Model of Surface to Air Missile Based on Differential Evolution Algorithm [J]. Journal of Air Force Engineering University:Natural Science ed,2009,10(5):41-44.
Systems Development Planning Model on Weapon System Under Uncertain Conditions
LIU Xu, LI Wei-min,SONG Wen-jing
(AFEU,Air and Missile Defense School,Shaanxi Xi’an 710051,China)
Abstract:The system development planning on weapon system under uncertain conditions is to study weapon system development planning model considering changing conditions. The systems development planning model of weapon system under certain conditions is built; the logistic function is introduced into the model to reflect the mapping relationship between investment funds and weapon capability. The uncertain conditions for the system development planning are explained with securities market risk investment theory. And the model under uncertain conditions is built based on Markowitz theory. The effectiveness of the proposed model and algorithm are demonstrated by a portfolio planning example of intelligence, surveillance, and reconnaissance weapon system.
Key words:uncertain conditions; weapon system of systems; development planning model; investment funds; weapons capability; Logistic; Markowitz
中圖分類號:E92;E917
文獻(xiàn)標(biāo)志碼:A
文章編號:1009-086X(2015)-05-0026-07
doi:10.3969/j.issn.1009-086x.2015.05.005
通信地址:710051陜西西安長樂東路甲字1號空軍工程大學(xué)防空反導(dǎo)學(xué)院研2隊E-mail:liuxu193@126.com
作者簡介:劉旭(1987-),男,湖南湘潭人。博士生,研究方向為作戰(zhàn)力量建設(shè)。
*收稿日期:2014-12-07;修回日期:2015-03-31