beta = params['beta'] # set stability parameter of levy flight mutation
etam = params['etam'] # set index parameter of polynomial mutation
#################
# start program #
#################
os.makedirs(f'./{prefix}',
exist_ok=True) # create a folder to include running results
os.makedirs(f'./{prefix}/history/', exist_ok=True)
os.makedirs(f'./{prefix}/history/{args.seed}', exist_ok=True)
W = das_dennis(n_part, n_obj) # generate a set of weight vectors
B = determine_neighbor(W, n_neb) # determine neighbor
X = init_pop(n_pop, n_var, xl, xu) # initialize a population
F = eval_pop(X, f) # evaluate fitness
z = init_ref_point(F) # determine a reference point
for c_gen in range(1, n_gen): # star main loop
result = np.hstack([F,
X]) # record objective values and decision variables
np.savetxt(f'./{prefix}/history/{args.seed}/{c_gen}.csv', result)
for i in np.random.permutation(n_pop): # traverse the population
xi, fi = X[i, :], F[i, :] # get current individual
if random.random() < sigma: # determine selection pool by probability
pool = B[i, :] # neighbor as the pool
else:
problem = set_problem(prob_name, n_var, n_obj, xu, xl) # set optimization problem ############################### # define MOEA/D configuration # ############################### W = get_weights(decomp_method, params) n_pop = len(W) params['n_pop'] = n_pop B = determine_neighbor(W, T) # determine neighbor X = init_pop(n_pop, n_var, xl, xu) # initialize a population Y = eval_pop(X, problem, prob_name) # evaluate fitness ref_point = init_ref_point(Y) # determine a reference point EP = init_EP(X, Y, n_pop, n_obj) #initialize External population beta = 0.3 P = np.zeros(n_pop) + beta # generate a set of stability beta parameters # P = np.random.uniform(params['betal'],params['betau'], n_pop) # generate a set of stability parameters nr = params['nr'] delta = params['delta'] ################################## # set self adaptive parameters # ################################## betal = params['betal'] # set lower bound for stability parameter of levy flight mutation
'beta for param'] # set stability factor of levy flight to search parameters
n_step = params['n_step'] # set number of generations to assess a parameter
#################
# start program #
#################
os.makedirs(f'./{output}',
exist_ok=True) # create a folder to include running results
os.makedirs(f'./{output}/history/', exist_ok=True)
os.makedirs(f'./{output}/history/{args.seed}', exist_ok=True)
W = das_dennis(sld_n_part, n_obj) # generate a set of weight vectors
B = determine_neighbor(W, T) # determine neighbor
X = init_pop(n_pop, n_var, xl, xu) # initialize a population
Y = eval_pop(X, problem) # evaluate fitness
z = init_ref_point(Y) # determine a reference point
P = np.random.uniform(betal, betau,
n_pop) # generate a set of stability parameters
P_parent = P[:] # initialize a list to store parameters in parent generations
P_offspring = np.full(
n_pop,
np.nan) # initialize a list to store parameters in offspring generations
I = np.zeros(n_pop) # initialize a set of indicator vaules
I_parent = I[:] # initialize a list to store indicators in parent generations
I_offspring = I[:] # initialize a list to store indicators in offspring generations
n_fe = n_pop