Commonly , crossover should have the ability to generate " better " individuals by not disrupting too much good patterns of the old ones . the power of gas arises from crossover . after we made a good investigation to some actual problems an efficient crossover , dislocation crossover ( dc ) , is proposed in this paper 交叉算子的設(shè)計(jì)和實(shí)現(xiàn)與所研究的問題密切相關(guān),所以應(yīng)該與個(gè)體編碼方案統(tǒng)一考慮,一般要求它既不要太多地破壞個(gè)體編碼串中表示優(yōu)良形狀的優(yōu)良模式,又要能夠有效地產(chǎn)生出一些較好的新個(gè)體模式。
In the second charper , two reformed metheds has presented , which are competition adaptation - ga based on elitist and dual mutation adaptive - ga which can increase the population diversity and can decrease the dependence of genetic algorithms " result to the control parameter of operators and the status of initial population . to testify their abilities of algorithms , some studies have excuted which included the study to optimize parameter for pid and the study to multivariable intelligent decoupling control for mimo system . the outcomes have showed us that this amendment has a better effective than conventional means and the genetic algorithms which have no change 第二章中,在已有文獻(xiàn)的基礎(chǔ)上,創(chuàng)新性的提出兩種改進(jìn)算法:基于最優(yōu)保留的聯(lián)賽競(jìng)爭(zhēng)機(jī)制遺傳算法( competitionadaptation - gabasedonelitist )和雙變異自適應(yīng)遺傳算法( dualmutationadaptive - ga ) ;通過對(duì)算法中選擇算子和變異算子的有效改進(jìn),提高了算法的尋優(yōu)能力和尋優(yōu)效率,增加了群體中個(gè)體模式的多樣性,對(duì)于算法中存在的欺騙問題、早熟問題以及成熟前收斂問題有明顯的改進(jìn)作用。