Source Code for Module Bio.GA.Selection.Diversity

 1  """Select individuals into a new population trying to maintain diversity. 
 2   
 3  This selection mechanism seeks to try and get both high fitness levels 
 4  and high diversity levels in the population. 
 5  """ 
 6  # standard modules 
 7  import random 
 8  import math 
 9   
10  # biopython 
11  from Bio.Seq import MutableSeq 
12   
13  # local modules 
14  from Abstract import AbstractSelection 
15  from Tournament import TournamentSelection 
16   
17   
18 -class DiversitySelection(AbstractSelection): 
19      """Implement diversity selection. 
20   
21      Diversity selection is performed by trying to select individuals 
22      from the population that aren't already in the new_population. A group 
23      of selected individuals is then subjected to selection using 
24      a passed selection routine. 
25   
26      If new individuals can not be selected, new individuals will be 
27      randomly generated and inserted into the population. 
28      """ 
29 -    def __init__(self, internal_selector, genome_generator): 
30          """Initialize a diversity selector. 
31   
32          Arguments: 
33   
34          o internal_selector - A selection object that will be used to select 
35          individuals based on fitness, perform crossover, mutation and repair. 
36   
37          o genome_generator - A function that, when called, will return a 
38          genome to be used for a new organism. The genome returned must 
39          be a MutableSeq() object. 
40          """ 
41          self._internal_selector = internal_selector 
42          self._genome_generator = genome_generator 
43   
44          self.sub_pop_percent = .1 
45          self.random_tries = 10 
46   
47 -    def _get_new_organism(self, new_pop, old_pop): 
48          """Get a new organism from old_pop that isn't in new_pop. 
49   
50          This attempts to select an organism from old_pop that isn't in 
51          new_pop. If we can't do this in the number of tries specified 
52          by the class attribute random_tries, we generate a new random 
53          organism and return that. 
54          """ 
55          # try to pick an organism that isn't in the population 
56          new_org = None 
57          num_tries = 0 
58          while new_org is None and num_tries < self.random_tries: 
59              chance_org = random.choice(old_pop) 
60   
61              if chance_org not in new_pop: 
62                  new_org = chance_org 
63   
64              num_tries += 1 
65   
66          # if we don't get an organism, generate a random one 
67          if new_org is None: 
68              new_org = old_pop[0].copy() 
69              random_genome = self._genome_generator() 
70              new_org.genome = random_genome 
71              new_org.recalculate_fitness() 
72   
73          return new_org 
74   
75 -    def select(self, population): 
76          """Perform selection on the current population, encouraging diversity. 
77          """ 
78          new_population = [] 
79   
80          while len(new_population) < len(population): 
81              # generate a sub population 
82              sub_pop_size = int(math.ceil(len(population) * 
83                                           self.sub_pop_percent)) 
84              sub_pop = [] 
85              for individual in range(sub_pop_size): 
86                  new_org = self._get_new_organism(new_population, population) 
87                  sub_pop.append(new_org) 
88   
89              # put the new sub population through selection, mutation 
90              # and all of that good stuff 
91              new_sub_pop = self._internal_selector.select(sub_pop) 
92   
93              new_population.extend(new_sub_pop) 
94   
95          # return the new population, which should have the same number 
96          # of individuals we started with. 
97          return new_population[:len(population)] 
98