MPI-Genetic-Algorithm/Project/worker.py at master · ivanprokopets/MPI-Genetic-Algorithm · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
import numpy as np
import numpy.random as rd
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.animation as animation
from multiprocessing.managers import BaseManager

from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
from collections import OrderedDict

from mpi4py import MPI
from time import time
from function import Func

comm = MPI.COMM_WORLD
rank = comm.Get_rank()
size = comm.Get_size()


import os


class WorkerGA:
    def __init__(self):
        self.data = {}
        self.data = comm.bcast(self.data, root=0)
        self.function = Func(self.data['function'])
        self.chromosomes = np.array([(2 * rd.rand(2) - 2) for i in range(self.data['chromosomes_number'])])
        all_gens = None
        all_values = None
        for i in range(self.data['generations_number']):
            self.chromosomes, values = self.next_generation(self.data['mutation'], self.data['optimizer'])
            if (i == 0):
                all_gens = self.chromosomes
                all_values = values
            else:
                all_gens = np.append(all_gens, self.chromosomes, axis=0)
                all_values = np.append(all_values, values, axis=0)
        recvbuf = None
        # if (rank ==  1):
            # print(all_gens[np.argsort(all_values)[0:4]])
        if(self.data['optimizer'] == 'min'):
            comm.Gather(all_gens[np.argsort(all_values)[0:4]], recvbuf, root=0)
        elif (self.data['optimizer'] == 'max'):
            comm.Gather(all_gens[np.argsort(all_values)[-3::]], recvbuf, root=0)


    def generate_new_part(self, chromosomes, mutation=False, optimizer='min'):
        values = [self.function(*chromosome) for chromosome in chromosomes]
        chromosomesDict = dict(zip([str(i) for i in range(4)], values))
        if optimizer == 'min':
            chromosomesDict = OrderedDict(sorted(chromosomesDict.items(), key=lambda t: t[1]))
        elif optimizer == 'max':
            chromosomesDict = OrderedDict(sorted(chromosomesDict.items(), key=lambda t: -t[1]))
        else:
            raise ValueError(str(optimizer) + 'should be max or min')

        chromosome_indexes = list()

        for chromosome_index in chromosomesDict.keys():
            chromosome_indexes.append(chromosome_index)

        good_chromosome = chromosomes[int(chromosome_indexes[2])]
        better_chromosome = chromosomes[int(chromosome_indexes[1])]
        best_chromosome = chromosomes[int(chromosome_indexes[0])]

        x0 = (better_chromosome[0] + good_chromosome[0]) / 2 + float(2 * mutation * rd.rand(1) - 1 * mutation)
        y0 = (better_chromosome[1] + 0.3 * good_chromosome[1]) + float(2 * mutation * rd.rand(1) - 1 * mutation)
        x1 = (better_chromosome[0] + best_chromosome[0]) / 2  + float(2 * mutation * rd.rand(1) - 1 * mutation)
        y1 = (better_chromosome[1] + best_chromosome[1]) / 2  + float(2 * mutation * rd.rand(1) - 1 * mutation)
        x2 = (best_chromosome[0] + good_chromosome[0]) / 2
        y2 = (best_chromosome[1] + good_chromosome[1]) / 2 + float(2 * mutation * rd.rand(1) - 1 * mutation)
        x3 = (better_chromosome[0] + 0.7 * good_chromosome[0]) / 1.7
        y3 = (best_chromosome[1] + 0.6 * better_chromosome[1]) / 1.6

        new_part = np.array([[x0, y0],
                             [x1, y1],
                             [x2, y2],
                             [x3, y3]])
        values = np.array([self.function(x0, y0), self.function(x1, y1), self.function(x2, y2), self.function(x3, y3)])

        return (new_part, values)

    def next_generation(self, mutation=False, optimizer='min'):
        part = np.array([self.chromosomes[j] for j in range(0, 4)])
        new_population, values = self.generate_new_part(part, mutation, optimizer)

        for parts_number in range(1, int(len(self.chromosomes) / 4)):
            part = np.array([self.chromosomes[j] for j in range(parts_number * 4, (parts_number + 1) * 4)])
            new_part, values_part = self.generate_new_part(part, mutation, optimizer)
            new_population = np.append(new_population, new_part, axis=0)
            values = np.append(values, values_part, axis=0)

        return (new_population, values)


    def calculate(self, optimizer='min'):
        if optimizer == 'min':
            return min([self.function(*chromosome) for chromosome in self.chromosomes])
        elif optimizer == 'max':
            return max([self.function(*chromosome) for chromosome in self.chromosomes])
        else:
            raise ValueError(optimizer + 'should be max or min')


if __name__ == '__main__':
    if (rank == 0):
        BaseManager.register('register_queue')
        BaseManager.register('input_queue')
        BaseManager.register('results_queue')
        m = BaseManager(address=('localhost', 8080), authkey=b'blah')
        m.connect()
        register_queue = m.register_queue()
        input_queue = m.input_queue()
        results_queue = m.results_queue()
        register_queue.put(size)
        print(rank, size)
        data = input_queue.get()
        print(data)

        comm.bcast(data, root=0)
        sendbuf = np.empty([data['chromosomes_number'], 2])
        recvbuf = np.empty([size, data['chromosomes_number'], 2])
        comm.Gather(sendbuf, recvbuf, root=0)
        chromosomes_to_send = None
        for i in range (1, size):
            if (i == 1):
                chromosomes_to_send = recvbuf[i]
            else:
                chromosomes_to_send = np.append(chromosomes_to_send, recvbuf[i], axis=0)
        results_queue.put(chromosomes_to_send)
    else:
        worker = WorkerGA()