Update to pygad.py to use multiprocessing of generations

pygad.py

@@ -4,6 +4,7 @@
 import pickle
 import time
 import warnings
+import concurrent.futures
 
 class GA:
 
@@ -47,7 +48,10 @@ def __init__(self,
                  save_best_solutions=False,
                  save_solutions=False,
                  suppress_warnings=False,
-                 stop_criteria=None):
+                 stop_criteria=None,
+                 use_multiprocess=False,
+                 max_workers=2
+                 ):
 
         """
         The constructor of the GA class accepts all parameters required to create an instance of the GA class. It validates such parameters.
@@ -105,6 +109,10 @@ def __init__(self,
         allow_duplicate_genes: Added in PyGAD 2.13.0. If True, then a solution/chromosome may have duplicate gene values. If False, then each gene will have a unique value in its solution.
 
         stop_criteria: Added in PyGAD 2.15.0. It is assigned to some criteria to stop the evolution if at least one criterion holds.
+
+        use_multiprocess: Added Dec 30, 2021: Whether or not to use concurrent.futures MultiprocessPoolExecutor to go through each chromosome of a population in a generation.
+
+        max_workers: Added Dec 30, 2021: How many cpu cores to use if "use_multiprocess" is set to "True"
         """
 
         # If suppress_warnings is bool and its valud is False, then print warning messages.
@@ -869,6 +877,12 @@ def __init__(self,
             self.valid_parameters = False
             raise TypeError("The expected value of the 'stop_criteria' is a single string or a list/tuple/numpy.ndarray of strings but the value {stop_criteria_val} of type {stop_criteria_type} found.".format(stop_criteria_val=stop_criteria, stop_criteria_type=type(stop_criteria)))
 
+        # Added warning for multiprocessing
+        self.use_multiprocess = use_multiprocess
+        self.max_workers = max_workers
+        if self.use_multiprocess == False and self.max_workers != 2:
+            if not self.suppress_warnings:  warnings.warn("You have designated a non-default number of workers, but 'use_multiprocess' is set to False. Multiprocessing will not be used in this run. Set 'use_multiprocess' to True if you want to use those workers that you set.")
+
         # The number of completed generations.
         self.generations_completed = 0
 
@@ -1131,6 +1145,54 @@ def initialize_population(self, low, high, allow_duplicate_genes, mutation_by_re
         # Keeping the initial population in the initial_population attribute.
         self.initial_population = self.population.copy()
 
+
+    def cal_pop_fitness_multiprocess(self):
+
+        """
+        Calculating the fitness values of all solutions in the current population, using multiprocessing. 
+        It returns:
+            -fitness: An array of the calculated fitness values.
+        """
+
+        if self.valid_parameters == False:
+            raise ValueError("ERROR calling the cal_pop_fitness_multiprocess() method: \nPlease check the parameters passed while creating an instance of the GA class.\n")
+
+        # Create a list of len(self.population) full of negative inf's
+        pop_fitness = [-numpy.inf] * len(self.population)
+
+        # First, do the same check if the parent's fitness value are already calculated. 
+        # If so, use it instead of calling the fitness function.
+        for sol_idx, sol in enumerate(self.population):
+            if not (self.last_generation_parents is None) and len(numpy.where(numpy.all(self.last_generation_parents == sol, axis=1))[0] > 0):
+                # Index of the parent in the parents array (self.last_generation_parents). This is not its index within the population.
+                parent_idx = numpy.where(numpy.all(self.last_generation_parents == sol, axis=1))[0][0]
+                # Index of the parent in the population.
+                parent_idx = self.last_generation_parents_indices[parent_idx]
+                # Use the parent's index to return its pre-calculated fitness value.
+                fitness = self.last_generation_fitness[parent_idx]
+                # NOTE: this is why we create the list ahead of time, so we can make use of the sol_idx
+                # for fitnesses already determined
+                pop_fitness[sol_idx] = fitness
+
+        # Start a process pool for each population member
+        with concurrent.futures.ProcessPoolExecutor(max_workers=self.max_workers) as executor:
+            # Don't do the ones we just checked above, based on the -numpy.inf in the pop_fitness list (again, using sol_idx)
+            # NOTE: Seeing as some fitnesses may finish quicker than others re: the process pool, I needed a way to still use
+            # the appropriate sol_idx for the particular chromosome being tested, and the easiest way I knew to do this was
+            # to have the fitness_func return the fitness, as well as the sol_idx that was passed into it. That way, regardless
+            # of which chromosome comes out of the fitness func first, it'll still be placed in its correct spot in the 
+            # pop_fitness list. So, if using multiprocess, the fitness_func must "return fitness, solution_idx"
+            results = [executor.submit(self.fitness_func, sol, sol_idx) for sol_idx, sol in enumerate(self.population) if pop_fitness[sol_idx] != -numpy.inf]
+            for f in concurrent.futures.as_completed(results):
+                solution_fitness = f.result()[0]
+                solution_idx = f.result()[1]
+                pop_fitness[solution_idx] = solution_fitness
+
+        # Same as default
+        pop_fitness = numpy.array(pop_fitness)
+
+        return pop_fitness
+
     def cal_pop_fitness(self):
 
         """
@@ -1184,7 +1246,10 @@ def run(self):
         stop_run = False
 
         # Measuring the fitness of each chromosome in the population. Save the fitness in the last_generation_fitness attribute.
-        self.last_generation_fitness = self.cal_pop_fitness()
+        if self.use_multiprocess == False:
+            self.last_generation_fitness = self.cal_pop_fitness()
+        elif self.use_multiprocess == True:
+            self.last_generation_fitness = self.cal_pop_fitness_multiprocess()
 
         best_solution, best_solution_fitness, best_match_idx = self.best_solution(pop_fitness=self.last_generation_fitness)
 
@@ -1258,7 +1323,10 @@ def run(self):
             self.generations_completed = generation + 1 # The generations_completed attribute holds the number of the last completed generation.
 
             # Measuring the fitness of each chromosome in the population. Save the fitness in the last_generation_fitness attribute.
-            self.last_generation_fitness = self.cal_pop_fitness()
+            if self.use_multiprocess == False:
+                self.last_generation_fitness = self.cal_pop_fitness()
+            elif self.use_multiprocess == True:
+                self.last_generation_fitness = self.cal_pop_fitness_multiprocess()
 
             best_solution, best_solution_fitness, best_match_idx = self.best_solution(pop_fitness=self.last_generation_fitness)
 
@@ -3111,8 +3179,12 @@ def best_solution(self, pop_fitness=None):
 
         # Getting the best solution after finishing all generations.
         # At first, the fitness is calculated for each solution in the final generation.
-        if pop_fitness is None:
-            pop_fitness = self.cal_pop_fitness()
+        if self.use_multiprocess == False:
+            if pop_fitness is None:
+                pop_fitness = self.cal_pop_fitness()
+        elif self.use_multiprocess == True:
+            if pop_fitness is None:
+                pop_fitness = self.cal_pop_fitness_multiprocess()
         # Then return the index of that solution corresponding to the best fitness.
         best_match_idx = numpy.where(pop_fitness == numpy.max(pop_fitness))[0][0]
 
@@ -3481,4 +3553,4 @@ def load(filename):
         raise FileNotFoundError("Error reading the file {filename}. Please check your inputs.".format(filename=filename))
     except:
         raise BaseException("Error loading the file. If the file already exists, please reload all the functions previously used (e.g. fitness function).")
-    return ga_in
+    return ga_in