def four_peaks(length, random_seeds):
# Define Four Peaks objective function and problem
four_fitness = FourPeaks(t_pct=0.1)
problem = DiscreteOpt(length=length,
fitness_fn=four_fitness,
maximize=True,
max_val=2)
problem.set_mimic_fast_mode(True) # set fast MIMIC
# Plot optimizations for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for SA, GA and MIMIC')
ro.plot_optimizations(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=100,
sa_max_iters=1000,
ga_max_iters=50,
mimic_max_iters=50,
sa_init_temp=100,
sa_decay_rates=np.arange(0.002, 0.04, 0.002),
sa_min_temp=0.001,
ga_pop_size=1000,
mimic_pop_size=1000,
ga_keep_pct=0.1,
mimic_keep_pct=0.2,
pop_sizes=np.arange(100, 1001, 100),
keep_pcts=np.arange(0.1, 0.51, 0.1),
plot_name='Four Peaks',
plot_ylabel='Fitness')
# Plot performances for RHC, SA, GA and MIMIC
print('\nPlot performances for RHC, SA, GA and MIMIC')
ro.plot_performances(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=1000,
sa_max_iters=5000,
ga_max_iters=100,
mimic_max_iters=100,
sa_init_temp=100,
sa_exp_decay_rate=0.02,
sa_min_temp=0.001,
ga_pop_size=1000,
ga_keep_pct=0.1,
mimic_pop_size=1000,
mimic_keep_pct=0.2,
plot_name='Four Peaks',
plot_ylabel='Fitness')
def Knapsack_f(length, weights, values, random_seeds):
Knapsack_objective = Knapsack(weights, values, max_weight_pct=0.35)
problem = DiscreteOpt(length=length,
fitness_fn=Knapsack_objective,
maximize=True,
max_val=2)
problem.set_mimic_fast_mode(True) # set fast MIMIC
# Plot optimizations for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for SA, GA and MIMIC')
ro.plot_optimizations(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=1000,
sa_max_iters=1000,
ga_max_iters=250,
mimic_max_iters=50,
sa_init_temp=100,
sa_decay_rates=np.arange(0.05, 2.01, 0.05),
sa_min_temp=0.001,
ga_pop_size=300,
mimic_pop_size=1500,
ga_keep_pct=0.2,
mimic_keep_pct=0.4,
pop_sizes=np.arange(100, 2001, 200),
keep_pcts=np.arange(0.1, 0.81, 0.1),
plot_name='Knapsack',
plot_ylabel='Fitness')
# Plot performances for RHC, SA, GA and MIMIC
print('\nPlot performances for RHC, SA, GA and MIMIC')
ro.plot_performances(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=1000,
sa_max_iters=1000,
ga_max_iters=500,
mimic_max_iters=100,
sa_init_temp=100,
sa_exp_decay_rate=0.020,
sa_min_temp=0.001,
ga_pop_size=100,
ga_keep_pct=0.2,
mimic_pop_size=300,
mimic_keep_pct=0.4,
plot_name='Knapsack',
plot_ylabel='Fitness')
def travel_salesman(length, distances, random_seeds):
# Define Travel Salesman objective function and problem
fitness_dists = TravellingSales(distances=distances)
# Define optimization problem object
problem = TSPOpt(length=length, fitness_fn=fitness_dists, maximize=False)
problem.set_mimic_fast_mode(True) # set fast MIMIC
# Plot optimizations for SA, GA and MIMIC
print('\nPlot Optimizations for SA, GA and MIMIC')
ro.plot_optimizations(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=500,
sa_max_iters=500,
ga_max_iters=50,
mimic_max_iters=50,
sa_init_temp=100,
sa_decay_rates=np.arange(0.005, 0.05, 0.005),
sa_min_temp=0.001,
ga_pop_size=100,
mimic_pop_size=100,
ga_keep_pct=0.2,
mimic_keep_pct=0.2,
pop_sizes=np.arange(100, 401, 100),
keep_pcts=np.arange(0.1, 0.51, 0.1),
plot_name='TSP',
plot_ylabel='Fitness')
#Plot performances for RHC, SA, GA and MIMIC
print('\nPlot Performances for RHC, SA, GA and MIMIC')
ro.plot_performances(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=500,
sa_max_iters=500,
ga_max_iters=500,
mimic_max_iters=500,
sa_init_temp=100,
sa_exp_decay_rate=0.02,
sa_min_temp=0.001,
ga_pop_size=100,
ga_keep_pct=0.15,
mimic_pop_size=300,
mimic_keep_pct=0.5,
plot_name='TSP',
plot_ylabel='Fitness')
def travel_salesman(length, distances, random_seeds):
"""Define and experiment the Travel Salesman optimization problem.
The Travel Salesman is a famous optimization problem presenting whose
purpose is to find the shortest tout of N cities and visit each exactly once.
Args:
length (int): problem length.
distances(list of tuples): list of inter-distances between each pair of cities.
random_seeds (list or ndarray): random seeds for get performances over multiple random runs.
Returns:
None.
"""
# Define Travel Salesman objective function and problem
tsp_objective = TravellingSales(distances=distances)
problem = TSPOpt(length=length, fitness_fn=tsp_objective, maximize=False)
problem.set_mimic_fast_mode(True) # set fast MIMIC
# Plot optimizations for SA, GA and MIMIC
print('\nPlot Optimizations for SA, GA and MIMIC')
ro.plot_optimizations(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=500, sa_max_iters=500, ga_max_iters=50, mimic_max_iters=10,
sa_init_temp=100, sa_decay_rates=np.arange(0.005, 0.05, 0.005), sa_min_temp=0.001,
ga_pop_size=100, mimic_pop_size=700, ga_keep_pct=0.2, mimic_keep_pct=0.2,
pop_sizes=np.arange(100, 1001, 100), keep_pcts=np.arange(0.1, 0.81, 0.1),
plot_name='TSP', plot_ylabel='Cost')
# Plot performances for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for RHC, SA, GA and MIMIC')
ro.plot_performances(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=500, sa_max_iters=500, ga_max_iters=50, mimic_max_iters=10,
sa_init_temp=100, sa_exp_decay_rate=0.03, sa_min_temp=0.001,
ga_pop_size=100, ga_keep_pct=0.2,
mimic_pop_size=700, mimic_keep_pct=0.2,
plot_name='TSP', plot_ylabel='Cost')
def four_peaks(length, random_seeds):
"""Define and experiment Four Peaks optimization problem.
Four Peaks is an optimization problem presenting 2 global optima
and 2 local optima.
Args:
length (int): problem length.
random_seeds (list or ndarray): random seeds for get performances over multiple random runs.
Returns:
None.
"""
# Define Four Peaks objective function and problem
four_fitness = FourPeaks(t_pct=0.1)
problem = DiscreteOpt(length=length, fitness_fn=four_fitness, maximize=True, max_val=2)
problem.set_mimic_fast_mode(True) # set fast MIMIC
# Plot optimizations for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for SA, GA and MIMIC')
ro.plot_optimizations(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=10000, sa_max_iters=10000, ga_max_iters=250, mimic_max_iters=250,
sa_init_temp=100, sa_decay_rates=np.arange(0.002, 0.1, 0.002), sa_min_temp=0.001,
ga_pop_size=1000, mimic_pop_size=1000, ga_keep_pct=0.1, mimic_keep_pct=0.2,
pop_sizes=np.arange(100, 1001, 100), keep_pcts=np.arange(0.1, 0.81, 0.1),
plot_name='Four Peaks', plot_ylabel='Fitness')
# Plot performances for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for RHC, SA, GA and MIMIC')
ro.plot_performances(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=7000, sa_max_iters=7000, ga_max_iters=250, mimic_max_iters=250,
sa_init_temp=100, sa_exp_decay_rate=0.02, sa_min_temp=0.001,
ga_pop_size=1000, ga_keep_pct=0.1,
mimic_pop_size=1000, mimic_keep_pct=0.2,
plot_name='Four Peaks', plot_ylabel='Fitness')
def flip_plop(length, random_seeds):
"""Define and experiment Flip Flop optimization problem.
Flip Flop is an optimization problem whose purpose is to optimize
the number of alternations in as vector of bits.
Args:
length (int): problem length.
random_seeds (list or ndarray): random seeds for get performances over multiple random runs.
Returns:
None.
"""
# Define Flip Flop objective function and problem
flip_flop_objective = FlipFlop()
problem = DiscreteOpt(length=length, fitness_fn=flip_flop_objective, maximize=True, max_val=2)
problem.set_mimic_fast_mode(True) # set fast MIMIC
# Plot optimizations for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for SA, GA and MIMIC')
ro.plot_optimizations(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=1000, sa_max_iters=1000, ga_max_iters=250, mimic_max_iters=50,
sa_init_temp=100, sa_decay_rates=np.arange(0.05, 2.01, 0.05), sa_min_temp=0.001,
ga_pop_size=300, mimic_pop_size=1500, ga_keep_pct=0.2, mimic_keep_pct=0.4,
pop_sizes=np.arange(100, 2001, 200), keep_pcts=np.arange(0.1, 0.81, 0.1),
plot_name='Flip Flop', plot_ylabel='Fitness')
# Plot performances for RHC, SA, GA and MIMIC
print('\nPlot Optimizations for RHC, SA, GA and MIMIC')
ro.plot_performances(problem=problem,
random_seeds=random_seeds,
rhc_max_iters=1000, sa_max_iters=1000, ga_max_iters=500, mimic_max_iters=100,
sa_init_temp=100, sa_exp_decay_rate=0.3, sa_min_temp=0.001,
ga_pop_size=300, ga_keep_pct=0.2,
mimic_pop_size=1500, mimic_keep_pct=0.4,
plot_name='Flip Flop', plot_ylabel='Fitness')