1 00:00:00,420 --> 00:00:07,260 Hello and welcome to this lecture, where we are going to implement a different function to select the 2 00:00:07,260 --> 00:00:13,830 parents in this part here off the flow of the genetic algorithm. 3 00:00:14,970 --> 00:00:25,470 And we will simulate this rule methods where the idea is to is being, they will in order to select 4 00:00:25,470 --> 00:00:30,360 some of the individuals, as we saw in last lecture. 5 00:00:30,690 --> 00:00:36,840 We have a higher probability of selecting goods solutions. 6 00:00:37,080 --> 00:00:45,960 However, there is a small probability off select the not that good solutions, as we can see here. 7 00:00:45,960 --> 00:00:54,630 For example, Ace three moving back to our Google collab fail, where we implement a new function in 8 00:00:54,630 --> 00:00:57,690 our genetic algorithm class. 9 00:00:58,740 --> 00:01:10,440 Let's define the function that select variants and it will receive as a parameter self, which is the 10 00:01:10,440 --> 00:01:11,730 default parameter. 11 00:01:12,030 --> 00:01:16,050 And also this some evaluation. 12 00:01:16,860 --> 00:01:28,650 The results of these function we implement is before it will return this some of this class of all individuals 13 00:01:28,650 --> 00:01:29,880 in the population. 14 00:01:30,180 --> 00:01:36,870 First, we will implement that, this function and then we will perform some tests. 15 00:01:37,110 --> 00:01:44,250 So it will be easier to understand how we can simulate the roulette methods. 16 00:01:44,700 --> 00:01:53,220 First, we will create a variable called Barings, which will be equal to minus one. 17 00:01:53,700 --> 00:02:03,680 These variable that we are going through returns, as we can see here, we have a list of individuals 18 00:02:03,960 --> 00:02:06,900 that are parts of the population. 19 00:02:07,200 --> 00:02:18,990 So the idea is to select one of these individuals here to apply crossover mutation to generate the new 20 00:02:18,990 --> 00:02:20,010 generation. 21 00:02:20,370 --> 00:02:27,720 So we will select a specific index of this list in our implementation. 22 00:02:27,960 --> 00:02:39,660 We create is a population composed of 20 individuals, so we will return an index which will be associate 23 00:02:39,660 --> 00:02:41,340 this to the population. 24 00:02:41,880 --> 00:02:51,450 This variable parents will be initialized with this value, which means that we are not to return any 25 00:02:51,450 --> 00:02:57,900 solo show because we know that indexes in Python starts at zero. 26 00:02:58,140 --> 00:03:03,630 So we need to select a number from zero to 19. 27 00:03:03,990 --> 00:03:12,190 Considering our population composed of 20 individuals, we will update this variable. 28 00:03:12,420 --> 00:03:16,020 Then we will create a new variable. 29 00:03:16,440 --> 00:03:29,610 Random value equals random and we will multiply by some evaluation the parameter that we are going to 30 00:03:29,610 --> 00:03:33,180 send just to perform some tests. 31 00:03:33,540 --> 00:03:42,660 We can go down here see that this evaluation equals one hundred and ninety six. 32 00:03:43,050 --> 00:03:49,590 So we are selecting a random number in this range when we run this code. 33 00:03:49,860 --> 00:03:58,590 We will get and the number in the range from zero to one and we can multiply by this sum. 34 00:03:59,730 --> 00:04:06,560 We will get our value in this range every time we run this code. 35 00:04:06,570 --> 00:04:18,690 We Gad's value one hundred sixty six to seven six to one, one hundred and thirteen. 36 00:04:18,690 --> 00:04:25,080 And so one, we will use this value to simulate that roulette. 37 00:04:25,650 --> 00:04:33,330 It will be this simulation when we spill the will to get one of the solutions. 38 00:04:33,900 --> 00:04:45,300 Let's go back to our glass and we will create another variable sum equals zero and I equals zero. 39 00:04:45,630 --> 00:04:50,160 The variable y will be used to control the loop. 40 00:04:50,580 --> 00:05:00,480 We are going to use a while loop while I less than length self population. 41 00:05:01,020 --> 00:05:09,240 It will guarantee that we are going to run these codes based on this size of the population. 42 00:05:09,710 --> 00:05:13,500 Will we implement a model condition and. 43 00:05:14,010 --> 00:05:25,590 Some less than they own, don't value, it will be easier to understand this condition when we passed 44 00:05:25,590 --> 00:05:31,320 this gold in some minutes, then we will increment this variable. 45 00:05:31,800 --> 00:05:37,730 Some blurs the equal cell population position. 46 00:05:37,740 --> 00:05:41,400 I thought score evaluation. 47 00:05:41,700 --> 00:05:46,830 We are doing something very similar what we implemented here. 48 00:05:47,100 --> 00:05:49,500 We are just adding these values. 49 00:05:50,010 --> 00:05:55,260 Then we will increments the parents plus equal one. 50 00:05:55,590 --> 00:05:59,310 And also I B+ equal one. 51 00:05:59,790 --> 00:06:08,400 Just a reminder that in this variable, we will return the index that was just selects. 52 00:06:08,670 --> 00:06:16,560 Or in other words, we will extracts an individual off these list here. 53 00:06:17,250 --> 00:06:23,010 Then we can just return after the while loop parents. 54 00:06:23,550 --> 00:06:30,300 And to better understand these codes, we can put some prints brands. 55 00:06:30,990 --> 00:06:39,920 Let's put some marks here to be easier to see the value random value when we run these codes. 56 00:06:39,930 --> 00:06:48,870 The first step is to generate a random value between zero and this sum evaluation. 57 00:06:49,200 --> 00:06:56,910 So let's put here the random value and inside the way you look. 58 00:06:57,540 --> 00:07:04,800 Let's variants first variable I and then this sum. 59 00:07:05,490 --> 00:07:10,430 And let's put here this sum will run again. 60 00:07:10,440 --> 00:07:13,350 This code should create the glass. 61 00:07:13,830 --> 00:07:20,440 Let's move back to this part where we are creating the new population. 62 00:07:21,060 --> 00:07:24,060 We can see the results here. 63 00:07:24,540 --> 00:07:28,770 The best in the video is in position zero. 64 00:07:29,460 --> 00:07:35,520 Then we can just run these codes to get the best in the video. 65 00:07:36,180 --> 00:07:46,260 We can prints this sum of évaluations equals eighteen thousand and just see the results. 66 00:07:46,590 --> 00:07:49,030 Let's create parents. 67 00:07:49,680 --> 00:07:53,520 One equals genetic algorithm. 68 00:07:53,940 --> 00:08:03,540 Let's call our new function select parents and we'll send the here this some less visualize parents. 69 00:08:04,170 --> 00:08:04,710 One. 70 00:08:05,700 --> 00:08:07,050 Run this code! 71 00:08:07,560 --> 00:08:17,070 See, that's the random value that was generated was one hundred and sixty seven. 72 00:08:17,430 --> 00:08:25,200 So we are in if the random value is less than this sum. 73 00:08:25,440 --> 00:08:35,700 As we can see here, these numbers here are equivalents to this sum of each one of the solutions. 74 00:08:36,270 --> 00:08:48,120 Take a look here that we keep odd the the values and then when I equals four, then we can take a look 75 00:08:48,120 --> 00:08:48,570 here. 76 00:08:48,900 --> 00:08:49,560 That's this. 77 00:08:49,560 --> 00:08:58,050 Some of the evaluations is less than the random value that was selected. 78 00:08:58,350 --> 00:09:08,070 So we stop is like, you think the codes and the results of the function is 14, which is the value 79 00:09:08,070 --> 00:09:10,770 of parents one variable. 80 00:09:11,070 --> 00:09:22,520 We can take a look here that we are selecting these individual words score is six thousands. 81 00:09:22,920 --> 00:09:27,270 Take a look here that we have bad their solutions. 82 00:09:27,540 --> 00:09:30,240 Should the initial parts of the list. 83 00:09:30,600 --> 00:09:40,380 But as I say adds in the theoretical lecture, the idea is to apply a random function to simulate that 84 00:09:40,380 --> 00:09:40,780 rule. 85 00:09:40,810 --> 00:09:41,880 Its methods. 86 00:09:42,450 --> 00:09:48,600 So we are not going to select only the best individuals. 87 00:09:48,990 --> 00:09:59,070 There is a chance of select the disordered individuals here that are not so good as these other ones 88 00:09:59,070 --> 00:09:59,610 here. 89 00:09:59,940 --> 00:10:09,060 But it is important to create diversity in this genetic algorithm, and the probability of selecting 90 00:10:09,300 --> 00:10:13,700 these last individuals are less than. 91 00:10:13,850 --> 00:10:21,890 Selecting these or other individuals in the initial positions off the list, because as we already know 92 00:10:22,250 --> 00:10:31,580 in the initial positions, we can see the best individuals, we can try to run this code again. 93 00:10:32,600 --> 00:10:36,790 See, that's another number was selected. 94 00:10:37,160 --> 00:10:48,740 Now the random value is 77 and doing the while loop, we are adding the scores of the individuals in 95 00:10:48,740 --> 00:10:50,240 the first position. 96 00:10:50,570 --> 00:10:52,220 That's conditional. 97 00:10:52,220 --> 00:10:55,100 Employment is in the way Loop is through. 98 00:10:55,490 --> 00:11:02,210 So we select this individual number five and then we can take a look here. 99 00:11:02,600 --> 00:11:04,040 12 thousands. 100 00:11:04,310 --> 00:11:14,660 And every time you run this code, you will get a different individual and it is based on this random 101 00:11:14,660 --> 00:11:17,330 value that is generated. 102 00:11:17,780 --> 00:11:26,510 Let's create another variable variance two equals genetic algorithm. 103 00:11:26,810 --> 00:11:28,520 Select variants. 104 00:11:28,730 --> 00:11:30,740 I will send again this sum. 105 00:11:31,340 --> 00:11:35,510 Let's visualize the index that was selected. 106 00:11:36,080 --> 00:11:45,260 See, that's in the first is equation with selected individual number four, and the SEC one is a commercial 107 00:11:45,560 --> 00:11:49,400 with selected individual number 12. 108 00:11:49,760 --> 00:12:01,700 So we can use both individuals to apply crossover and mutation to generate the new generation. 109 00:12:02,240 --> 00:12:07,970 So now we have implemented this select variants function. 110 00:12:08,270 --> 00:12:18,380 We will also select two random variants and then we are going to apply the genetic operators. 111 00:12:18,680 --> 00:12:23,300 That's what we are going to implement in the next lectures. 112 00:12:23,700 --> 00:12:24,380 So there.