王瑩
摘要:針對(duì)引力搜索算法求解精度不高,易于早熟等缺點(diǎn),提出了一種改進(jìn)的引力搜索算法。為了平衡算法的開發(fā)與探索能力,我們引入兩個(gè)變異算子:一個(gè)算子增強(qiáng)算法的開發(fā)能力;另一算子增強(qiáng)算法的探索能力。最后把改進(jìn)算法應(yīng)用到典型測(cè)試題中,數(shù)值結(jié)果表明,算法是可行的,有效的。
Abstract: This paper proposes an improved gravitational search algorithm aiming at the shortcomings of the gravitational search algorithm, such as low precision and easy prematurity. In order to balance the development and exploration capabilities of the algorithm, we introduce two mutation operators: one operator enhances the ability to develop the algorithm; the other operator enhances the ability to explore the algorithm. Finally, the improved algorithm is applied to typical test questions. Numerical results show that the algorithm is feasible and effective.
關(guān)鍵詞:引力搜索算法;算法改進(jìn);變異算子
Key words: gravity search algorithm;algorithm improvement;mutation operator
中圖分類號(hào):TP18 文獻(xiàn)標(biāo)識(shí)碼:A 文章編號(hào):1006-4311(2018)21-0234-03
0 引言
引力搜索算法(Gravitational Search Algorithm,GSA)是Esmat Rashedi等人受萬(wàn)有引力定律和牛頓運(yùn)動(dòng)學(xué)第二定律啟發(fā)在2009年提出的一種新興的啟發(fā)式優(yōu)化算法。自從引力搜索算法被提出以來(lái),它已被廣泛地應(yīng)用于實(shí)際生活中,如神經(jīng)網(wǎng)絡(luò)訓(xùn)練[1],軟件工程[2],圖像處理[3]和動(dòng)力工程[4]等諸多問(wèn)題,引力搜索算法顯然已成為解決優(yōu)化問(wèn)題的一種十分重要的算法。一些學(xué)者提出許多改進(jìn)的引力搜索算法[5]-[7]。為了平衡引力搜索算法的開發(fā)與探索能力,本文提出一種改進(jìn)的引力搜索算法。在該算法中,引入兩個(gè)變異算子,一個(gè)算子具有開發(fā)能力另一算子具有探索能力。從而克服引力搜索算法收斂快,易于早熟的缺點(diǎn)。
1 引力搜索算法
3 數(shù)值實(shí)驗(yàn)
為了評(píng)價(jià)算法的性能,我們選取5個(gè)測(cè)試函數(shù)分別是Sphere(F1),Schwefel's2.22(F2),Schwefel's2.21(F3),Generalized Rastrigin(F4),Ackley(F5)針對(duì)每個(gè)問(wèn)題兩個(gè)算法在MATLB 2007環(huán)境下獨(dú)立運(yùn)行30次,所得結(jié)果見(jiàn)表1。兩種算法的參數(shù)設(shè)置如下:β=20,G0=100,最大迭代步數(shù)tmax=1000。從表1看出IOGSA在F1和F4上優(yōu)于GSA,所以IOGSA算法是可行且有效的。
4 結(jié)論
引力搜索算法雖然有很強(qiáng)的全局搜索能力,但是在計(jì)算的后期卻缺乏開發(fā)能力。本論文的總體目標(biāo)就是有效地平衡了算法的探索能力和開發(fā)能力得到更加有效的改進(jìn)的引力搜索算法IOGSA。
參考文獻(xiàn):
[1]Beatriz González a, Fevrier Valdez a, Patricia Melina,*, German Prado-Arechiga b, Fuzzy logic in the gravitational search algorithm enhanced using fuzzy logic with dynamic alpha parameter value adaptation for the optimization of modular neural networks in echocardiogram recognition, Applied Soft Computing, Applied Soft Computing,2015,37,245-254.
[2]Palanikkumar D., ANBUSELVAN P. and RITHU B. A Gravitational Search Algorithm for effective Web Service Selection for Composition with enhanced QoS in SOA, International Journal of Computer Applications [J]. 2012, 42(8): 12-15.
[3]Zhao W.Adaptive Image Enhancement Based on Gravitational Search Algorithm [J].Procedia Engineering, 2011,15:3288-3292.
[4]Mood, S. E., E. Rasshedi, M. M. Javidi. New Functions for Mass Caculation in Gravitational Search Algorithm. - Journal of Computing and Security, 2016,2.
[5]張維平,任雪飛,李國(guó)強(qiáng),等.改進(jìn)的萬(wàn)有引力搜索算法在函數(shù)優(yōu)化中的應(yīng)用[J].計(jì)算機(jī)應(yīng)用,2013,33(5):1317-1320.
[6]Rashedi E., Nezamabadi-pour H., Saryazdi S., BGSA: binary gravitational search algorithm, Natural Computing,2009,9,727-745.
[7] Li P., Duan H. B., Path planning of unmanned aerial vehicle based on improved gravitational search algorithm[J]. Science China Technological Sciences, 2012, 55, 2712-2719.