Merge branch 'plotScore' into 'main'

add a function to plot the average score

See merge request ii/collabschool22/genetic-bike-a!15

main.py

@@ -2,9 +2,10 @@
 
 import src.genetic_algo as ga
 import src.simu_bike as sb
+import src.scorePlot as sp
 
 gen=ga.Genetic_algorithm(population_size=15,
                          start_point=(0,5),
                          side_length=3,
                          fit_func=sb.creat_bike)
-gen.run()
+sp.plot(gen.run())
\ No newline at end of file

src/bike_no_animation.py

-import src.gen as g
-import numpy as np
-import sys
 import matplotlib.pyplot as plt
-from matplotlib.animation import FuncAnimation
 
-# initial, set up the physical world
-# the positions of points are generated using random function from np
-num_point = 4
-#points_position = np.random.rand(num_point, 2)*8+3
-
-# for test
-points_position = np.array([[1.0, 5],
-                             [5, 5],
-                             [1, 10],
-                             [5, 10]])
-points_position+=5
-start_x = np.max(points_position[:, 0])
-points_velocity = np.zeros_like(points_position)
-
-points_mass = np.array([20, 20, 0.1, 0.1])  # unit [m]
-wheel_radius = 1.2  # unit [m]
-
-spring_indexs = []
-for i in range(num_point):
-    for j in range(num_point):
-        if i != j:
-            spring_indexs.append([i, j])
-spring_indexs = np.array(spring_indexs)
-
-spring_stiffness = 1e3  # unit [N/m] a.k.a -> K
-
-gravity = -9.81  # unit [ m /s2]
-spring_initial_length = g.calc_spring_length(points_position, spring_indexs)
-
-# for animation
-run_time = 15  # unit [s]
-dt = 0.01     # unit [s]
-
-fig, ax = plt.subplots()
-ax.set_title('Bike animation')
-ax.axis('equal')
-ax.set_xlim(0, 35)
-ax.set_ylim(-5, 20)
-
-ax.plot(np.arange(-10, 40, 0.5), g.ground(np.arange(-10, 40, 0.5)), 'k')  # ground
-
-balls = []
-balls.append(plt.Circle(
-    (points_position[0][0], points_position[0][1]), wheel_radius, color='r'))
-balls.append(plt.Circle(
-    (points_position[1][0], points_position[1][1]), wheel_radius))
-
-lines = []
-for item in spring_indexs:
-    lines.append(ax.plot([points_position[item[0], 0], points_position[item[1], 0]],
-                         [points_position[item[0], 1], points_position[item[1], 1]],
-                         'o-', markersize=9, markerfacecolor='b'))
-for item in balls:
-    ax.add_patch(item)
-
-
-def update(frame):
-    global points_position, points_velocity
-    # boundary condition 1
-    if np.min(points_position[2:, 1] - g.ground(points_position[2:, 0])) < 0:
-        print('Max-distance is %.2f m' % (np.max(points_position[:, 0])-start_x))
-        sys.exit(0)
-    else:
-        ax.set_title('Bike animation at %.2f s\n Max-distance is %.2f m' %
-                     (frame*dt, np.max(points_position[:, 0])-start_x))
-
-        state_y = points_position[:, 1] - g.ground(points_position[:, 0])
-        state_y[0] -= wheel_radius
-        state_y[1] -= wheel_radius
-
-        # update position
-        points_position += points_velocity*dt
-
-        # update velocity
-        spring_length = g.calc_spring_length(points_position, spring_indexs)
-        spring_force = g.calc_spring_force(
-            spring_length-spring_initial_length, spring_stiffness)
-
-        points_velocity += g.calc_point_acceleration(
-            points_position, spring_indexs, spring_force, points_mass)*dt
-        points_velocity[:,
-                        1] += g.ground_force(state_y, spring_stiffness, points_mass)*dt
-        points_velocity[:, 1] += gravity*dt
-
-        # damp the vertical velocity to make the solution stable
-        yfliter = state_y * points_velocity[:, 1] > 0
-        points_velocity[:, 1][yfliter] *= 0.01
-        # boundary condition 2
-        if points_position[0, 1] < wheel_radius+g.ground(points_position[0, 0]):
-            points_velocity[0, 0] = 2
-
-        # rendering
-        for index, item in enumerate(spring_indexs):
-            lines[index][0].set_data([points_position[item[0], 0], points_position[item[1], 0]],
-                                     [points_position[item[0], 1], points_position[item[1], 1]],)
-        for i in [0, 1]:
-            balls[i].center = (points_position[i][0], points_position[i][1])
-
-        return lines, balls
-
-
-#ani = FuncAnimation(
-#    fig,
-#    update,
-#    np.arange(1, int(run_time/dt)),  # the length of simulation
-#    interval=10,
-#    repeat=False,
-#)
-
-
-#plt.show()
-
-for i in np.arange(1, int(run_time/dt)):
-    update(i)
-print(np.max(points_position[:, 0])-start_x)
+def plot(scores):
+    plt.plot(range( len(scores)),scores,'--b')
+    plt.show()
\ No newline at end of file

src/genetic_algo.py

@@ -287,6 +287,7 @@ random.randint(start_point[1], start_point[1]+side_length)),
         best_fitness_score = self.p.best_fitness_score()
         # REPEAT  
         file_counter=0
+        score_list=[]
         while abs(best_fitness_score - average_fitness_score) > 0.5:
             self.p.save_population_to_file("data/population_"+str(file_counter)+".json")
             file_counter+=1
@@ -302,7 +303,10 @@ random.randint(start_point[1], start_point[1]+side_length)),
             print(f'{average_fitness_score=}, {best_fitness_score=}')                     
             # UNTIL population has converged
             # STOP
-
+            score_list.append(average_fitness_score)
+            if len(score_list)>50:
+                break
+        return score_list
         #return the fittest child among the last siblings
         # remaining_sibling = sorted(p.population, key=lambda x: x.fitness_score)
         # remaining_sibling.reverse()         

src/scorePlot.py

+import matplotlib.pyplot as plt
+
+def plot(scores):
+    plt.plot(range( len(scores)),scores,'--b')
+    plt.show()