Preliminary study of genetic algorithm

Wrote a 10 number of the maximum sum, the value range of what 0-99, because the state space is relatively large, so the probability of mutation selection of 0.3, larger, crossover probability 0.8, population size 100, genetic algebra 1000

effect, not very ideal, distance 990 of the limit there is no small gap

As can be seen, only 8 times above 950, and after the increase of the genetic algebra to 10000, the effect did not increase significantly, quite failed

#include <bits/stdc++.h>using namespacestd;Const Doubleby=0.3;//mutation ProbabilityConst DoubleJh=0.8;//Free exchange probabilityConst intMax_size= -;Const intMax_d=100000;intsum;structnode{intv[Ten]; intFitness; voidCol () {Fitness=0;  for(intI=0;i<Ten; i++) Fitness+=V[i]; }}now[ the],max,min;//The population, the temporary, the largest individual, the smallest individual .//Generate initial populationvoidinit () { for(intI=0; i<max_size;i++){         for(intj=0;j<Ten; j + +) Now[i].v[j]=rand ()% -; }}voidjs () {sum=0; Max.fitness=-1; min.fitness=0x3f3f3f3f;  for(intI=0; i<max_size;i++) {now[i].col (); if(now[i].fitness>max.fitness) max=Now[i]; if(now[i].fitness<min.fitness) min=Now[i]; Sum+=now[i].fitness; }}//Print an individualvoidprint_o (Node A) { for(intI=0;i<Ten; i++) printf ("%d", A.v[i]); printf ("%d\n", a.fitness);}//Print all current stocksvoidprint () { for(intI=0; i<max_size;i++) Print_o (Now[i]);}intrand_id () {intD=rand ()% (sum+1); intCnt=0;  for(intI=0; i<max_size;i++) {CNT+=now[i].fitness; if(Cnt>=d)returni; }    returnRAND ()%max_size;}Doublerandd () {DoubleAns=rand ()% +; Ans/= +; returnans;}voidTran () {JS (); //find the best individuals    inttmp=0, I,j,pos;  for(intI=1; i<max_size;i++)        if(now[i].fitness>now[tmp].fitness) TMP=i;  for(intk=0; k<max_size;k+=2){        //Select a cross-entityI=rand_id (); J=rand_id (); //Intersection PositionPos=rand ()%Ten; //Exchange        if(RandD () <JH)            {swap (Now[k].v[pos],now[j].v[pos]);        Swap (Now[k].v[pos],now[i].v[pos]); }        //mutation        if(RandD () <y) {POS=rand ()%Ten; Now[k].v[pos]=rand ()% -;    }} JS (); intId=0;  for(intI=1; i<max_size;i++)      if(now[id].fitness>now[i].fitness) ID=i; Now[id]=Max;}voidga () {intCnt=0;  while(cnt<max_d) {tran (); CNT++; //print (); //puts ("\ n");    }    intans=now[0].fitness;  for(intI=0; i<max_size;i++) ans=Max (ans,now[i].fitness); printf ("%d\n", ans);}intMain () {srand (unsigned) time (NULL));  for(intI=1; i<= the; i++) {printf ("Case %d: \ n", i);        Init ();    GA (); }    return 0;}

View Code

Preliminary study of genetic algorithm