求助，如何把遗传算法接进一个优化调配过程，谢谢

Question

这是一个优化问题的目标函数，
function modeobjectfun(f1:single;f2:single;f3:single):single;//求解目标函数的模； f1为目标函数1，f2为目标函数2  f3为目标函数3
   begin
      Result:=Sqrt(f1*f1+f2*f2+f3*f3);
   end;这是我网上下载的一个遗传算法，请问如何接近程序里？
unit Unit1;{
计算
f(x1,x2)=100*(x1^2-x2)^2+(1-x1)^2的
/***************************************************************/
*          Simple Genetic Algorithm              
*                  2002/7/18
*
*         对于整体过程的说明
*     main
*            GenerateInitialPopulation //产生初始子代，本程序产生0，1 序列 
*            EvalutePopulation  
*                   |
*    |--------CalculateObjectValue    //计算目标函数值
*    |
*    |--------CalculateFitnessValue   //计算适应度
*    | 
*    |--------FindBestAndWorstIndividual//计算最好与最差个体
*   
*        以下循环到GENERATION代      
*            GenerateNextPopulation
*                   |
*    |--------SelectOperator            //选择操作
*                   |
*                   |--------CalculateFitnessValue     //单点交叉
*                   |
*                   |--------FindBestAndWorstIndividual //变异操作
*            
*            EvalutePopulation   //该函数同上
*                          .......
*                          .......
*                          .......
*            PerformEvolution         //遗传
*            OutputTextReport         //输出报告
*        循环结束    
****************************************************************
}
interfaceuses
  Windows, Messages, SysUtils, Variants, Classes, Graphics, Controls, Forms,
  Dialogs, StdCtrls;const
  CPopSize = 160 ;          //Simple Genetic Algorithm
  //基因长度
  CLENGTH1 = 11 ;           //第一个变量的遗传变量的长度
  CLENGTH2 = 11 ;           //第二个变量的遗传变量的长度
  CCHROMLENGTH = CLENGTH1 + CLENGTH2 ;   //变量总长度  CMAXIMIZATION = 1 ;       //最大标志
  CMINIZATION = 2;         //最小标志  CCmax=100;
  CCmin = 0;  CPc=0.6;               //交叉概率
  CPm=0.05;             //变异概率
  CMaxGeneration=1500;    //最大遗传代数type
   individual = record     //个体结构
   chrom: array[1..CCHROMLENGTH] of char;
   value: double;
   fitness: double;
end;type TGIndividual = array[1..CPopSize] of individual;type
  TForm1 = class(TForm)
    Button2: TButton;
    ListBox1: TListBox;
    procedure Button2Click(Sender: TObject);
  private
    { Private declarations }
  public
    { Public declarations }
    procedure generateinitpopulation(PopSize,CHROMLENGTH,MAXIMIZATION: integer;
        var functionmode: integer;var population: TGIndividual);
    procedure Init();
    function DecodeChromosome(s: string;strstart,strend: integer):longint;
    procedure calculateobjectvalue(PopSize,LENGTH1,LENGTH2: integer;
                var population: TGIndividual);
    procedure calculatefitnessvalue(PopSize,MAXIMIZATION,MINIZATION,CMIN,Cmax: integer;
                var functionmode: integer;var population: TGIndividual);
    procedure EvaluatePopulation;
    procedure generatenextpopulation;
    procedure selectionoperator(PopSize: integer;
                 var population: TGIndividual);
  //  procedure crossoveroperator(PopSize,Pc: integer;
    //             var population: TGIndividual);
    procedure crossoveroperator(PopSize,CHROMLENGTH: integer; Pc: double;
                var population: TGIndividual);    procedure findbestandworstindividual(PopSize: integer;var bestindividual,worstindividual,currentbest: individual;
                 var population: TGIndividual;var best_index,worst_index,generation: integer);
    procedure mutationoperator(PopSize,CHROMLENGTH: integer; pm: double;
               var population: TGIndividual);
    procedure performevolution(bestindividual: individual;var currentbest: individual;
                 best_index,worst_index: integer; var population: TGIndividual);
    procedure GeneticAlgorithmMain();
    procedure outputtextreport;
  end;var
  Form1: TForm1;
  g_functionmode: integer;//最优化类型
  g_population: TGIndividual;//种群空间
  g_bestindividual: individual;       //当前代最优秀个体
  g_worstindividual: individual;       //当前代最差个体
  g_generation: integer;
  g_currentbest: individual;       //现在得到的最优秀个体
  g_best_index: integer;          //记录最优秀个体
  g_worst_index: integer;          //记录最劣质个体
implementation{$R *.dfm}//生成最先一代
procedure TForm1.generateinitpopulation(PopSize,CHROMLENGTH,MAXIMIZATION: integer;
        var functionmode: integer;var population: TGIndividual);
var
  i,j:integer; 
begin
   randomize;
   for i:=1 to PopSize do                   //种群大小   
     for j:=1 to CHROMLENGTH do             //基因长度
       if random(10)<5 then
         population[i].chrom[j]:='0'    //以相同概率产生'0','1'
       else
         population[i].chrom[j]:='1';
   functionMode := MAXIMIZATION;               //最优化类型
   functionMode := CMINIZATION ;
end;

Answer 1

接着上面
procedure TForm1.Init();
begin
  //生成新一代
  generateinitpopulation(CPopSize,CCHROMLENGTH,CMAXIMIZATION,g_functionmode,g_population);
end;procedure TForm1.EvaluatePopulation;
begin
  //计算每个个体的函数值
  calculateobjectvalue(CPopSize,CLENGTH1,CLENGTH2,g_population);
  //计算每个个体的适应度
  calculatefitnessvalue(CPopSize,CMAXIMIZATION,CMINIZATION,CCMIN,CCmax,
               g_functionmode,g_population);
  //找到最好的个体和最差的个体
  findbestandworstindividual(CPopSize,g_bestindividual,g_worstindividual,g_currentbest,
                 g_population,g_best_index,g_worst_index,g_generation);
end;procedure TForm1.generatenextpopulation;
begin
  //按概率选择新的个体，产生新的一代
  mutationoperator(CPopSize,CCHROMLENGTH,Cpm,g_population);
  crossoveroperator(CPopSize,CCHROMLENGTH,CPc,g_population);
  selectionoperator(CPopSize,g_population);
end;function TForm1.DecodeChromosome(s: string;strstart,strend: integer):longint;
//基因解码过程,对于本程序,是将2进制基因转化成10进制数值
var
    decimal: longint;
    temp: longint;
    i: integer;
begin
   decimal:=0;
   temp:=1;    //以左到右，权值升高
   for i:=strstart to strend do
     begin
       if copy(s,i,1)='1' then
         decimal:=decimal+temp;   //转化到十进制       temp:=temp*2;
     end;
   decodechromosome:=decimal;
end;procedure TForm1.calculateobjectvalue(PopSize,LENGTH1,LENGTH2: integer;
                var population: TGIndividual);
//计算目标函数值,本程序目前计算为f(x1,x2)=100*(x1^2-x2)^2+(1-x1)^2
var
    i:integer;
    temp1,temp2:longint;
    x1,x2:double;
begin
   for i:=1 to PopSize do
     begin
       temp1:=DecodeChromosome(population[i].chrom,1,LENGTH1);
                 //分别对两个基因解码
       temp2:=DecodeChromosome(population[i].chrom,LENGTH1+1,LENGTH1+LENGTH2);       //从这里可以看出,本程序是采用均匀量化
       x1:=4.096*temp1/1023.0-2.048;
       x2:=4.096*temp2/1023.0-2.048;//       population[i].value := 100*(x1*x1-x2)*(x1*x1-x2)+(1-x1)*(1-x1);
       population[i].value := 10.2 + (x1 * x1 - 10 * cos (2 * pi * x1 ))
                              + ( x2 - 10 * cos (2 * pi * x2 ) );
    //   listbox1.Items.Add(floattostr(x1));
    //   listbox1.Items.Add(floattostr(x2));
     //  population[i].value := ( x1 + x2 ) * ( x1 + x2 ) - 4 * ( x1 + x2 ) + 8 ;
  //     listBox1.Items.Add(floattostr( x1 + x2 ));
     end;
end;procedure TForm1.calculatefitnessvalue(PopSize,MAXIMIZATION,MINIZATION,CMIN,Cmax: integer;
    var functionmode: integer;var population: TGIndividual);
//计算适应度
var
    i: integer;
    temp: double;
begin
   temp := 0 ;
   for i := 1 to PopSize do
   begin
       if functionmode = MAXIMIZATION then
       begin
         //常数Cmin := 0 
         if population[i].value + Cmin > 0.0  then
            temp := Cmin+population[i].value
         else
            temp := 0.0 ;
       end       else
       begin
        if functionmode = MINIZATION then
          if population[i].value < Cmax then
            temp := Cmax-population[i].value
          else
            temp := 0.0;
       end;
       
       population[i].fitness := temp;
   end;
end;procedure TForm1.findbestandworstindividual(PopSize: integer;var bestindividual,worstindividual,currentbest: individual;
                 var population: TGIndividual;var best_index,worst_index,generation: integer);
//找到当前种群中最优秀个体
var
   i: integer;
   sum: double;
begin
   sum:=0.0;
   bestindividual := population[1];
   worstindividual := population[1];
   
   for i:=2 to PopSize do
   begin
     //找到当前种群中最优秀的个体
     if population[i].fitness>bestindividual.fitness then
     begin
       bestindividual := population[i];
       best_index := i;
     end;     //找到当前种群中最差的个体
     if population[i].fitness<worstindividual.fitness then
     begin
       worstindividual := population[i];
       worst_index := i;
     end;
     //得到总的适应度
     sum:=sum+population[i].fitness;
   end;
   
   if generation = 1 then
     currentbest := bestindividual
   else
     if bestindividual.fitness > currentbest.fitness then
       currentbest := bestindividual;
end;procedure TForm1.selectionoperator(PopSize: integer;
              var population: TGIndividual);
//通过比例选择产生新的基因
var
   i,index,j: integer;
   p,sum: double;
   cfitness:array [1..CPopSize] of double;
   newpopulation: TGIndividual;
begin  
   sum := 0.0;   //求得上一代适应度的总和
   for i := 1 to PopSize do
     sum := sum + population[i].fitness;   //求得上一代每个个体的的适应度占总适应度的百分比
   for i := 1 to PopSize do
   begin
     if sum = 0 then
       cfitness[i] := 0
     else
       cfitness[i] := population[i].fitness / sum;
   end;
   //使得每一个适应度为前面的累加  
   for i := 2 to PopSize do
     cfitness[i] := cfitness[i-1] + cfitness[i];    //按概率得到新的种群 
    for i := 1 to popsize do
    begin
       Randomize ;
       p := random(1000) / 1000.0;
       
       index := 1;
       while p > cfitness[index] do
        index := index+1;       newpopulation[i] := population[index];
    end;    //将新的种群覆盖久的种群
    for i := 1 to popsize do
      population[i]:=newpopulation[i];
end;procedure TForm1.crossoveroperator(PopSize,CHROMLENGTH: integer; Pc: double;
                var population: TGIndividual);
//单点交叉
var
    i,j:integer;
    index:array[1..CPopSize] of integer;
    point,temp:integer;
    p:double;
    ch:char;
begin
   Randomize;   for i:=1 to PopSize do
     index[i]:=i;
     
   for i:=1 to PopSize do
   begin
       point:=random(PopSize-i+1);
       temp:=index[i];
       index[i]:=index[point+i];
       index[point+i]:=temp;
   end;
   //随机选择一对个体,单点交叉
   i := 1 ;
   while i <= PopSize do
   begin
     p:=random(1000) / 1000.0;     if p<Pc then
     begin
       point:=random(CHROMLENGTH-1)+1;
       for j:=point to CHROMLENGTH do
       begin
         ch:=population[index[i]].chrom[j];
         population[index[i]].chrom[j]:=population[index[i+1]].chrom[j];
         population[index[i+1]].chrom[j]:=ch;
         end;
     end;
     i := i + 2 ;
   end;
end;procedure TForm1.mutationoperator(PopSize,CHROMLENGTH: integer; pm: double;
               var population: TGIndividual);
//基因变异
var
   i,j: integer;
   p: double;
begin
   for i := 1 to PopSize do
    for j := 1 to CHROMLENGTH do
      begin
        p:=random(1000) / 1000.0; //这里控制随机发生变异
        if p < Pm then
        begin
          if population[i].chrom[j] = '0' then
            population[i].chrom[j] := '1'
          else population[i].chrom[j] := '0';
        end;
      end;
end;procedure TForm1.performevolution(bestindividual: individual;var currentbest: individual;
                 best_index,worst_index: integer; var population: TGIndividual);
//进化
begin
   if bestindividual.fitness>currentbest.fitness then
     currentbest:=population[best_index]
   else
     population[worst_index] := currentbest;
end;procedure TForm1.GeneticAlgorithmMain();
begin
  g_generation := 1;
  EvaluatePopulation;
  while g_generation < CMaxGeneration do
  begin
    g_generation := g_generation + 1;
    generatenextpopulation;
    EvaluatePopulation;
    performevolution(g_bestindividual,g_currentbest,g_best_index,g_worst_index,g_population);
  end;  outputtextreport;
end;procedure TForm1.outputtextreport();
var
   i,j:integer;
   str: string;
begin
   for j:=1 to Cpopsize do
   begin
      str := 'NO.' + inttostr(j);
      Str := Str + '  ' + '基因1：';
      for i:=1 to Clength1 do
         Str := str + g_currentbest.chrom[i] + ' ' ;
      Str := Str + '  ' + '基因2：';
      for i:=clength1+1 to clength1+clength2 do
         Str := Str + g_currentbest.chrom[i] + ' ' ;
      Str := Str + '  ' + ' 适应度：';
      Str := Str + FloatToStr( g_currentbest.fitness );      Str := Str + '  ' + ' 函数值：';
      Str := Str + FloatToStr( g_currentbest.value );
      ListBox1.Items.Add(Str);
   end;
   ListBox1.Items.Add(inttostr(g_generation));
end;
procedure TForm1.Button2Click(Sender: TObject);
begin
  ListBox1.Items.Add('begin');
  Init();
  GeneticAlgorithmMain();
  ListBox1.Items.Add('end');
end;end.