هل يمكنك اعادة كتابة كود بايثون التالي بالماتلاب مع مراعاة الفهارس في ماتلاب تبدأ ب 1…
- User
هل يمكنك اعادة كتابة كود بايثون التالي بالماتلاب مع مراعاة الفهارس في ماتلاب تبدأ ب 1 بدل 0 #!/usr/bin/env python3
-- coding: utf-8 --
""" Created on Thu Sep 13 13:23:01 2018
@author: krupa """
=============================================================================
INITIAL ENCODING GA
1. For each particle record:
a. Route list [[cust1, cust20,...]....[cust7,cust80...]]
b Distance list [Dr1, .......Drk]
c distance
d num_routes
2. record fitness table : distance, num_routes, fitness function value
=============================================================================
#%%
import os import math import multiprocessing import numpy as np import pandas as pd import random from initial_pop_generation import initial_population from initial_routing_phase2 import initial_routing_phase2 #%% #Paths cwd = os.getcwd() customer_path = cwd +'/solomon25_csv/customers' dataInfo_path = cwd +'/solomon25_csv/data_info' distanceMatrix_path = cwd +'/solomon25_csv/distance_matrix' initial_path = cwd + '/initial'
#%%
LIST FILE NAMES
r1_list_datasets = [] c1_list_datasets = [] rc1_list_datasets = [] r2_list_datasets = [] c2_list_datasets = [] rc2_list_datasets = []
for i in range(101,110): c1_list_datasets.append('c'+'%s' %(i) )
for i in range(101,113): r1_list_datasets.append('r'+'%s' %(i) )
for i in range(101,109): rc1_list_datasets.append('rc'+'%s' %(i) )
for i in range(201,209): c2_list_datasets.append('c'+'%s' %(i) )
for i in range(201,212): r2_list_datasets.append('r'+'%s' %(i) )
for i in range(201,209): rc2_list_datasets.append('rc'+'%s' %(i) )
dataset_list = c1_list_datasets + r1_list_datasets + rc1_list_datasets + c2_list_datasets + r2_list_datasets + rc2_list_datasets
r1_list_dataInfo = [] c1_list_dataInfo = [] rc1_list_dataInfo = [] r2_list_dataInfo = [] c2_list_dataInfo = [] rc2_list_dataInfo = []
for i in range(101,110): c1_list_dataInfo.append('c'+'%s' %(i) + 'dataInfo.csv')
for i in range(101,113): r1_list_dataInfo.append('r'+'%s' %(i) + 'dataInfo.csv')
for i in range(101,109): rc1_list_dataInfo.append('rc'+'%s' %(i) + 'dataInfo.csv')
for i in range(201,209): c2_list_dataInfo.append('c'+'%s' %(i) + 'dataInfo.csv')
for i in range(201,212): r2_list_dataInfo.append('r'+'%s' %(i) + 'dataInfo.csv')
for i in range(201,209): rc2_list_dataInfo.append('rc'+'%s' %(i) + 'dataInfo.csv')
dataInfo_files_list = c1_list_dataInfo + r1_list_dataInfo + rc1_list_dataInfo + c2_list_dataInfo + r2_list_dataInfo + rc2_list_dataInfo
#customers r1_list_customers = [] c1_list_customers = [] rc1_list_customers = [] r2_list_customers = [] c2_list_customers = [] rc2_list_customers = []
for i in range(101,110): c1_list_customers.append('c'+'%s' %(i) + 'customers.csv')
for i in range(101,113): r1_list_customers.append('r'+'%s' %(i) + 'customers.csv')
for i in range(101,109): rc1_list_customers.append('rc'+'%s' %(i) + 'customers.csv')
for i in range(201,209): c2_list_customers.append('c'+'%s' %(i) + 'customers.csv')
for i in range(201,212): r2_list_customers.append('r'+'%s' %(i) + 'customers.csv')
for i in range(201,209): rc2_list_customers.append('rc'+'%s' %(i) + 'customers.csv')
customers_files_list = c1_list_customers + r1_list_customers + rc1_list_customers + c2_list_customers + r2_list_customers + rc2_list_customers
#distance matrix #customers r1_list_distanceMatrix = [] c1_list_distanceMatrix = [] rc1_list_distanceMatrix = [] r2_list_distanceMatrix = [] c2_list_distanceMatrix = [] rc2_list_distanceMatrix = []
for i in range(101,110): c1_list_distanceMatrix.append('c'+'%s' %(i) + 'distanceMatrix.csv')
for i in range(101,113): r1_list_distanceMatrix.append('r'+'%s' %(i) + 'distanceMatrix.csv')
for i in range(101,109): rc1_list_distanceMatrix.append('rc'+'%s' %(i) + 'distanceMatrix.csv')
for i in range(201,209): c2_list_distanceMatrix.append('c'+'%s' %(i) + 'distanceMatrix.csv')
for i in range(201,212): r2_list_distanceMatrix.append('r'+'%s' %(i) + 'distanceMatrix.csv')
for i in range(201,209): rc2_list_distanceMatrix.append('rc'+'%s' %(i) + 'distanceMatrix.csv')
distanceMatrix_files_list = c1_list_distanceMatrix + r1_list_distanceMatrix + rc1_list_distanceMatrix + c2_list_distanceMatrix + r2_list_distanceMatrix + rc2_list_distanceMatrix
#write out files
r1_list_initialPop = [] c1_list_initialPop = [] rc1_list_initialPop = [] r2_list_initialPop = [] c2_list_initialPop = [] rc2_list_initialPop = []
for i in range(101,110): c1_list_initialPop.append('c'+'%s' %(i) + 'initialPop.npy')
for i in range(101,113): r1_list_initialPop.append('r'+'%s' %(i) + 'initialPop.npy')
for i in range(101,109): rc1_list_initialPop.append('rc'+'%s' %(i) + 'initialPop.npy')
for i in range(201,209): c2_list_initialPop.append('c'+'%s' %(i) + 'initialPop.npy')
for i in range(201,212): r2_list_initialPop.append('r'+'%s' %(i) + 'initialPop.npy')
for i in range(201,209): rc2_list_initialPop.append('rc'+'%s' %(i) + 'initialPop.npy')
initialPop_files_list = c1_list_initialPop + r1_list_initialPop + rc1_list_initialPop + c2_list_initialPop + r2_list_initialPop + rc2_list_initialPop
r1_list_initialDistance = [] c1_list_initialDistance = [] rc1_list_initialDistance = [] r2_list_initialDistance = [] c2_list_initialDistance = [] rc2_list_initialDistance = []
for i in range(101,110): c1_list_initialDistance.append('c'+'%s' %(i) + 'initialDistance.npy')
for i in range(101,113): r1_list_initialDistance.append('r'+'%s' %(i) + 'initialDistance.npy')
for i in range(101,109): rc1_list_initialDistance.append('rc'+'%s' %(i) + 'initialDistance.npy')
for i in range(201,209): c2_list_initialDistance.append('c'+'%s' %(i) + 'initialDistance.npy')
for i in range(201,212): r2_list_initialDistance.append('r'+'%s' %(i) + 'initialDistance.npy')
for i in range(201,209): rc2_list_initialDistance.append('rc'+'%s' %(i) + 'initialDistance.npy')
initialDistance_files_list = c1_list_initialDistance + r1_list_initialDistance + rc1_list_initialDistance + c2_list_initialDistance + r2_list_initialDistance + rc2_list_initialDistance
r1_list_initialResults = [] c1_list_initialResults = [] rc1_list_initialResults = [] r2_list_initialResults = [] c2_list_initialResults = [] rc2_list_initialResults = []
for i in range(101,110): c1_list_initialResults.append('c'+'%s' %(i) + 'initialResults.csv')
for i in range(101,113): r1_list_initialResults.append('r'+'%s' %(i) + 'initialResults.csv')
for i in range(101,109): rc1_list_initialResults.append('rc'+'%s' %(i) + 'initialResults.csv')
for i in range(201,209): c2_list_initialResults.append('c'+'%s' %(i) + 'initialResults.csv')
for i in range(201,212): r2_list_initialResults.append('r'+'%s' %(i) + 'initialResults.csv')
for i in range(201,209): rc2_list_initialResults.append('rc'+'%s' %(i) + 'initialResults.csv')
initialResults_files_list = c1_list_initialResults + r1_list_initialResults + rc1_list_initialResults + c2_list_initialResults + r2_list_initialResults + rc2_list_initialResults
#%% #EXPERIMENTAL VARIABLES num_experiments = 30 num_cores = multiprocessing.cpu_count() #number of cores to use
#GLOBAL VARIABLES: N = 300#population size greedy_percent = 0.10#percentage to be greedily initialised random_percent = 0.90#percentage to be randomily initialised alpha = 100 beta = 0.001
#%% i = 0 experiment = 0 #%% #FOR EACH DATASET for i in range(len(dataset_list)): dataSet = dataset_list[i] #READ IN DATA PERTAINING TO THE DATASET customer_file = os.path.join(customer_path, customers_files_list[i]) dataInfo_file = os.path.join(dataInfo_path, dataInfo_files_list[i]) distanceMatrix_file = os.path.join(distanceMatrix_path, distanceMatrix_files_list[i])
df_customers = pd.read_csv(customer_file) df_distance_matrix = pd.read_csv(distanceMatrix_file) df_dataInfo = pd.read_csv(dataInfo_file) num_customers = df_customers.shape[0] num_customers_depot = num_customers + 1 df_customers.index = range(1,num_customers_depot) number_random_chromosomes = math.floor(N*random_percent)-1 number_nearest_nn_chromosomes = 1 number_greedy_chromosomes = N-number_random_chromosomes-1#int(N*greedy_percent) #CONVERT TO NUMPY ARRAY #distance array arr_distance_matrix = np.asarray(df_distance_matrix) #greedy radius dataset_radius = (np.nanmax(arr_distance_matrix) - np.nanmin(arr_distance_matrix))/2 #customer arraycustomer_index = 0
demand = 1
service_time = 2
ready_time = 3
due_time = 4
complete_time = 5
arr_customers = np.empty((num_customers, 6)) arr_customers[:,0] = np.arange(1, num_customers_depot) arr_customers[:,1] = df_customers.loc[:,'demand'] arr_customers[:,2] = df_customers.loc[:,'serviceTime'] arr_customers[:,3] = df_customers.loc[:,'readyTime'] arr_customers[:,4] = df_customers.loc[:,'dueTime'] arr_customers[:,5] = df_customers.loc[:,'completeTime']fleet max details
total_time = df_dataInfo.loc[0, 'fleet_max_working_time'] total_capacity = df_dataInfo.loc[0, 'fleet_capacity'] all_customers_list = list(range(1,num_customers_depot))FOR EACH EXPERIMENT
for experiment in range(num_experiments):READ OUT FILES
initial_pop_output_file = os.path.join(initial_path + '/%s/experiment%d' %(dataSet,experiment), initialPop_files_list[i]) initial_pop_distance_output_file = os.path.join(initial_path + '/%s/experiment%d' %(dataSet,experiment), initialDistance_files_list[i]) initial_result_output_file = os.path.join(initial_path + '/%s/experiment%d' %(dataSet,experiment), initialResults_files_list[i]) radius = dataset_radius*(random.uniform(0.5,1)) #assumed NOTE: NOT STATED IN PAPER #phase1 pop_chromosome_routeList_array, pop_chromosome_distanceList_array, pop_num_routes, pop_total_distance = initial_population(N, num_cores, number_random_chromosomes, number_greedy_chromosomes, number_nearest_nn_chromosomes, num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list, radius) #phase2pop_chromosome_routeList_array, pop_chromosome_distanceList_array, pop_num_routes, pop_total_distance = initial_population_phase2(N, num_cores, arr_customers, arr_distance_matrix, total_capacity, total_time, pop_chromosome_routeList_array, pop_chromosome_distanceList_array, pop_num_routes, pop_total_distance, alpha, beta)
pop_chromosome_routeList_array, pop_chromosome_distanceList_array, arr_pop_results = initial_routing_phase2(num_cores,N, alpha, beta, pop_chromosome_routeList_array, pop_chromosome_distanceList_array, arr_distance_matrix, arr_customers, total_time, total_capacity) np.save(initial_pop_output_file, pop_chromosome_routeList_array) np.save(initial_pop_distance_output_file, pop_chromosome_distanceList_array) df_initial_results = pd.DataFrame(np.zeros((N, 3))) df_initial_results.columns = ['num_vehicles', 'distance', 'fitness'] df_initial_results.loc[:, 'num_vehicles'] = pop_num_routes df_initial_results.loc[:, 'distance'] = pop_total_distance #METRIC: GA PAPER alpha*numV + beta*distance df_initial_results.loc[:,'fitness'] = (alpha*pop_num_routes)+(beta*pop_total_distance) df_initial_results.to_csv(initial_result_output_file, index=False) print('experiment: ', experiment) print('Dataset:', dataSet)#!/usr/bin/env python3
-- coding: utf-8 --
""" Created on Fri Sep 14 13:32:53 2018
@author: krupa """
=============================================================================
GENERATE THE INITIAL POPULATION ROUTES USING THE 3 SCHEMES;
1. RANDOM
2. GREEDY
=============================================================================
from initial_routing import random_neighbour, greedy_neighbour, nearest_neighbour import concurrent.futures as cf from itertools import chain import numpy as np #%%
def initial_population(N, num_cores, number_random_chromosomes, number_greedy_chromosomes, number_nearest_nn_chromosomes, num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list, radius): with cf.ProcessPoolExecutor(num_cores) as executor: results_list = [] for i in range(number_random_chromosomes): results_list.append(executor.submit(random_neighbour, num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list)) for j in range(number_greedy_chromosomes): results_list.append(executor.submit(greedy_neighbour, num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list, radius))
results_list.append(nearest_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list))
chromosome_route_list, chromosome_distance_list, num_routes, total_distance
temp = list(chain.from_iterable(f.result() for f in cf.as_completed(results_list))) pop_chromosome_routeList_array = np.empty(N, object) pop_chromosome_distanceList_array = np.empty(N, object) pop_num_routes = np.empty(N) pop_total_distance = np.empty(N) chromosome_route_list, chromosome_distance_list, num_routes, total_distance = nearest_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list) pop_chromosome_routeList_array[0] = chromosome_route_list pop_chromosome_distanceList_array[0] = chromosome_distance_list pop_num_routes[0] = num_routes pop_total_distance[0] = total_distance pop_chromosome_routeList_array[1::] = temp[0::4] pop_chromosome_distanceList_array[1::] = temp[1::4] pop_num_routes[1::] = temp[2::4] pop_total_distance[1::] = temp[3::4] return pop_chromosome_routeList_array, pop_chromosome_distanceList_array, pop_num_routes, pop_total_distance#!/usr/bin/env python3
-- coding: utf-8 --
""" Created on Fri Sep 14 14:41:23 2018
@author: krupa
""" #%%
=============================================================================
PHASE II ROUTING:
1. REMOVE LAST CUSTOMER OF R1
2. ADD LASTE CUSTOMER TO BE FIRST CUSTOME OF R2
3. IF BOTH FEASIBLE AND NEW SUM DISTANCE IS LESS OR REDUCES NUMBER OF VEHICLES THEN ACCEPT
=============================================================================
#%% #import copy #import collections from validity import route_validity import numpy as np from joblib import Parallel, delayed
#%%
def routing_phase2(alpha, beta, chromosome, chromosome_distance_list, total_capacity, total_time, arr_distance_matrix, arr_customers): index_list = np.arange(0, len(chromosome)+1) index_list[len(chromosome)] = 0 for i in range(len(index_list)-1): r1_index = index_list[i] r2_index = index_list[i+1] route_r1 = chromosome[r1_index].copy() route_r2 = chromosome[r2_index].copy() dist_r1 = chromosome_distance_list[r1_index] dist_r2 = chromosome_distance_list[r2_index] customer_r1 = route_r1[-2] route_r2.insert(1, customer_r1) route_validity_status, new_r2_dist = route_validity(route_r2, arr_distance_matrix, arr_customers, total_time, total_capacity) if route_validity == True: if len(route_r1)>3: new_r1_dist = dist_r1 - arr_distance_matrix[route_r1[-2],0]+ arr_distance_matrix[route_r1[-3],0] if (dist_r1+dist_r2)>(new_r1_dist+new_r2_dist): del route_r1[-1] chromosome[r1_index] = route_r1 chromosome[r2_index] = route_r2 chromosome_distance_list[r1_index] = new_r1_dist chromosome_distance_list[r2_index] = new_r2_dist else: del chromosome[r1_index] chromosome[r2_index] = route_r2 del chromosome_distance_list[r1_index] chromosome_distance_list[r2_index] = new_r2_dist total_distance = sum(chromosome_distance_list) num_vehicles = len(chromosome) fitness = (alphanum_vehicles)+(betatotal_distance) return chromosome, chromosome_distance_list, num_vehicles, total_distance, fitness
#%%
def initial_routing_phase2(num_cores,N, alpha, beta, pop_chromosome_routeList_array, pop_chromosome_distanceList_array, arr_distance_matrix, arr_customers, total_time, total_capacity): inputs = range(0,N) results_list = Parallel(num_cores)(delayed(routing_phase2)(alpha, beta, pop_chromosome_routeList_array[i], pop_chromosome_distanceList_array[i], total_capacity, total_time, arr_distance_matrix, arr_customers ) for i in inputs)
chromosome, chromosome_distance_list, num_vehicles, total_distance, fitness
unzipped = zip(*results_list) #groups the same return for each particle and returns a list of tuples each of lenth M results_list = list(unzipped) pop_chromosome_routeList_array = np.empty(N, object) pop_chromosome_distanceList_array = np.empty(N, object) pop_chromosome_routeList_array[:] = results_list[0] pop_chromosome_distanceList_array[:] = results_list[1] arr_pop_results = np.empty((N,3), float) arr_pop_results[:,0] = results_list[2] arr_pop_results[:,1] = results_list[3] arr_pop_results[:,2] = results_list[4] return pop_chromosome_routeList_array, pop_chromosome_distanceList_array, arr_pop_results#!/usr/bin/env python3
-- coding: utf-8 --
""" Created on Thu Sep 13 17:04:34 2018
@author: krupa """ from feasible_neighbours import feasible_neighbours import random #%%
=============================================================================
INITIAL ROUTING SCHEME
1. Random
2. Greedy
3. Phase 2
=============================================================================
#%% def random_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list): available_customers_list = all_customers_list.copy()
chromosome_route_list = [] chromosome_distance_list = [] route_list = [0] curr_capacity = 0 curr_time = 0 curr_customer = 0 curr_distance = 0 while len(available_customers_list)!=0: arr_feasible_neighbours = feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, available_customers_list, total_time, total_capacity, curr_time, curr_capacity, curr_customer) if arr_feasible_neighbours.shape[0]!=0:#has neighbours, append them feasible_neighbours_list = arr_feasible_neighbours[:,0].tolist() curr_customer = int(random.choice(feasible_neighbours_list)) curr_customer_index = feasible_neighbours_list.index(curr_customer) route_list.append(curr_customer) curr_distance += arr_feasible_neighbours[curr_customer_index, 6] curr_capacity = arr_feasible_neighbours[curr_customer_index, 1] curr_time = arr_feasible_neighbours[curr_customer_index, 10] available_customers_list.remove(curr_customer) else: #end route and initialise a new route route_list.append(0) chromosome_route_list.append(route_list) curr_distance += arr_distance_matrix[curr_customer,0] chromosome_distance_list.append(curr_distance) #initialise new route route_list = [0] curr_capacity = 0 curr_time = 0 curr_customer = 0 curr_distance = 0 #end last route route_list.append(0) chromosome_route_list.append(route_list) curr_distance += arr_distance_matrix[curr_customer,0] chromosome_distance_list.append(curr_distance) num_routes= len(chromosome_distance_list) total_distance = sum(chromosome_distance_list) random_route_results = (chromosome_route_list, chromosome_distance_list, num_routes, total_distance) return random_route_results#%% def greedy_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list, radius): available_customers_list = all_customers_list.copy()
chromosome_route_list = [] chromosome_distance_list = [] route_list = [0] curr_capacity = 0 curr_time = 0 curr_customer = 0 curr_distance = 0 while len(available_customers_list)!=0: arr_feasible_neighbours = feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, available_customers_list, total_time, total_capacity, curr_time, curr_capacity, curr_customer) if arr_feasible_neighbours.shape[0]!=0:#has neighbours, append them if arr_feasible_neighbours[arr_feasible_neighbours[:,6]<=radius].shape[0] !=0: arr_feasible_neighbours = arr_feasible_neighbours[arr_feasible_neighbours[:,6]<=radius] #reduce search space to feasible neighbours withn radius feasible_neighbours_list = arr_feasible_neighbours[:,0].tolist() curr_customer = int(random.choice(feasible_neighbours_list)) curr_customer_index = feasible_neighbours_list.index(curr_customer) route_list.append(curr_customer) curr_distance += arr_feasible_neighbours[curr_customer_index, 6] curr_capacity = arr_feasible_neighbours[curr_customer_index, 1] curr_time = arr_feasible_neighbours[curr_customer_index, 10] available_customers_list.remove(curr_customer) else: #end route and initialise a new route route_list.append(0) chromosome_route_list.append(route_list) curr_distance += arr_distance_matrix[curr_customer,0] chromosome_distance_list.append(curr_distance) #initialise new route route_list = [0] curr_capacity = 0 curr_time = 0 curr_customer = 0 curr_distance = 0 #end last route route_list.append(0) chromosome_route_list.append(route_list) curr_distance += arr_distance_matrix[curr_customer,0] chromosome_distance_list.append(curr_distance) num_routes= len(chromosome_distance_list) total_distance = sum(chromosome_distance_list) greedy_route_results = (chromosome_route_list, chromosome_distance_list, num_routes, total_distance) return greedy_route_results#%% def nearest_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list): available_customers_list = all_customers_list.copy()
chromosome_route_list = [] chromosome_distance_list = [] route_list = [0] curr_capacity = 0 curr_time = 0 curr_customer = 0 curr_distance = 0 while len(available_customers_list)!=0: arr_feasible_neighbours = feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, available_customers_list, total_time, total_capacity, curr_time, curr_capacity, curr_customer) if arr_feasible_neighbours.shape[0]!=0:#has neighbours, append them #find minimum travel distance curr_customer_index = arr_feasible_neighbours[:,6].argmin() curr_customer = int(arr_feasible_neighbours[curr_customer_index,0] ) route_list.append(curr_customer) curr_distance += arr_feasible_neighbours[curr_customer_index, 6] curr_capacity = arr_feasible_neighbours[curr_customer_index, 1] curr_time = arr_feasible_neighbours[curr_customer_index, 10] available_customers_list.remove(curr_customer) else: #end route and initialise a new route route_list.append(0) chromosome_route_list.append(route_list) curr_distance += arr_distance_matrix[curr_customer,0] chromosome_distance_list.append(curr_distance) #initialise new route route_list = [0] curr_capacity = 0 curr_time = 0 curr_customer = 0 curr_distance = 0 #end last route route_list.append(0) chromosome_route_list.append(route_list) curr_distance += arr_distance_matrix[curr_customer,0] chromosome_distance_list.append(curr_distance) num_routes= len(chromosome_distance_list) total_distance = sum(chromosome_distance_list) return chromosome_route_list, chromosome_distance_list, num_routes, total_distance#!/usr/bin/env python3
-- coding: utf-8 --
""" Created on Thu Sep 13 16:08:53 2018
@author: krupa """ #%% import numpy as np
#%%
=============================================================================
FEASIBLE NEIGHBOURS
copy array of all customers from arr_customers
to arr_feasible_customers with the following columns
customer_index = 0
demand = 1
service_time = 2
ready_time = 3
due_time = 4
complete_time = 5
distance_curr_next = 6
arrival_time = 7
waiting_time = 8
start_time = 9
finish_time = 10
return_time = 11
=============================================================================
#%% def feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, available_customers_list, total_time, total_capacity, curr_time, curr_capacity, curr_customer): arr_feasible_customers = np.empty((len(available_customers_list), 12)) available_customers_list_index = available_customers_list.copy() available_customers_list_index[:] = [x - 1 for x in available_customers_list_index] arr_feasible_customers[:,0:6] = arr_customers[available_customers_list_index, :]
#curr capacity arr_feasible_customers[:, 1] = arr_feasible_customers[:, 1] + curr_capacity #distance_curr_next arr_feasible_customers[:, 6] = arr_distance_matrix[curr_customer, available_customers_list] #arrival_time arr_feasible_customers[:, 7] = arr_feasible_customers[:, 6] + curr_time #waiting_time arr_feasible_customers[:, 8] = arr_feasible_customers[:, 3] - arr_feasible_customers[:, 7] arr_feasible_customers[:, 8] = arr_feasible_customers[:, 8].clip(min=0) #start time arr_feasible_customers[:, 9] = arr_feasible_customers[:, 7] + arr_feasible_customers[:, 8] #arrival + waiting #finish_time arr_feasible_customers[:, 10] = arr_feasible_customers[:, 9] + arr_feasible_customers[:, 2] #(arrival + waiting )+ service #return_time arr_feasible_customers[:, 11] = arr_distance_matrix[available_customers_list, 0] + arr_feasible_customers[:, 10] #keep only those that adhere to constraints arr_feasible_customers = arr_feasible_customers[(arr_feasible_customers[:,1]<=total_capacity) & (arr_feasible_customers[:,11]<=total_time) & (arr_feasible_customers[:,9]>=arr_feasible_customers[:,3]) & (arr_feasible_customers[:,10]<=arr_feasible_customers[:,5])] return arr_feasible_customers#!/usr/bin/env python3
-- coding: utf-8 --
""" Created on Mon Sep 17 10:46:08 2018
@author: krupa """
#import numpy as np #%%
=============================================================================
Given a route [depot, c1, ....cn, 0]
check validity of the route: time, capacity constraints.
want to see if the sequence of the route is valid
=============================================================================
#%% def route_validity(route, arr_distance_matrix, arr_customers, total_time, total_capacity): route_validity_status = True curr_dist = arr_distance_matrix[route[0], route[1]] curr_capacity = arr_customers[route[1]-1, 1] #demand of customer (route[1]) -1 index postion ready_time = arr_customers[route[1]-1, 3] due_time = arr_customers[route[1]-1, 4] curr_time = arr_distance_matrix[route[0], route[1]] waiting_time = max(0, ready_time-curr_time) curr_time += waiting_time + arr_customers[route[1]-1, 2] #service time
route_customers = route[1:-1] for i in range(len(route_customers)-1): curr_dist += arr_distance_matrix[route_customers[i], route_customers[i+1]] curr_time += arr_distance_matrix[route_customers[i], route_customers[i+1]] ready_time = arr_customers[route_customers[i+1]-1, 3] due_time = arr_customers[route_customers[i+1]-1, 4] service_time = arr_customers[route_customers[i+1]-1, 2] curr_capacity += arr_customers[route_customers[i+1]-1, 1] if curr_time<=due_time and curr_capacity<total_capacity: #accept waiting_time = max(0, ready_time - curr_time) if (curr_time + waiting_time>=ready_time) and (curr_time + waiting_time + service_time <= total_time): #accept next customer curr_time += waiting_time + service_time else: #route invalid route_validity_status = False break else: #route invalid route_validity_status = False break if route_validity_status == True: curr_dist += arr_distance_matrix[route[-2], route[-1]] curr_time += arr_distance_matrix[route[-2], route[-1]] if curr_time>total_time: route_validity_status = False return route_validity_status, curr_dist - Assistant
Translating the provided Python code to MATLAB involves several steps, mainly due to differences in syntax, programming paradigms, and built-in functions between the two languages. Additionally, MATLAB uses 1-based indexing, which requires careful adjustments throughout the code.
Below is a comprehensive translation of your Python scripts into MATLAB. The translation includes:
- Main Script (
main.m): Corresponds to your primary Python script. - Helper Functions: Each Python script is translated into a separate MATLAB function:
initial_population.minitial_routing_phase2.mrouting_phase2.mrandom_neighbour.mgreedy_neighbour.mnearest_neighbour.mfeasible_neighbours.mroute_validity.m
1. Main Script (
main.m)% main.m % MATLAB translation of the provided Python script % Clear workspace and command window clear; clc; % Define paths cwd = pwd; customer_path = fullfile(cwd, 'solomon25_csv', 'customers'); dataInfo_path = fullfile(cwd, 'solomon25_csv', 'data_info'); distanceMatrix_path = fullfile(cwd, 'solomon25_csv', 'distance_matrix'); initial_path = fullfile(cwd, 'initial'); % Initialize dataset lists c1_list_datasets = arrayfun(@(x) sprintf('c%d', x), 101:109, 'UniformOutput', false); r1_list_datasets = arrayfun(@(x) sprintf('r%d', x), 101:112, 'UniformOutput', false); rc1_list_datasets = arrayfun(@(x) sprintf('rc%d', x), 101:108, 'UniformOutput', false); c2_list_datasets = arrayfun(@(x) sprintf('c%d', x), 201:208, 'UniformOutput', false); r2_list_datasets = arrayfun(@(x) sprintf('r%d', x), 201:211, 'UniformOutput', false); rc2_list_datasets = arrayfun(@(x) sprintf('rc%d', x), 201:208, 'UniformOutput', false); dataset_list = [c1_list_datasets, r1_list_datasets, rc1_list_datasets, ... c2_list_datasets, r2_list_datasets, rc2_list_datasets]; % Generate file lists with extensions generate_file_list = @(prefix, range, suffix) arrayfun(@(x) sprintf('%s%d%s', prefix, x, suffix), range, 'UniformOutput', false); dataInfo_files_list = [generate_file_list('c', 101:109, 'dataInfo.csv'), ... generate_file_list('r', 101:112, 'dataInfo.csv'), ... generate_file_list('rc', 101:108, 'dataInfo.csv'), ... generate_file_list('c', 201:208, 'dataInfo.csv'), ... generate_file_list('r', 201:211, 'dataInfo.csv'), ... generate_file_list('rc', 201:208, 'dataInfo.csv')]; customers_files_list = [generate_file_list('c', 101:109, 'customers.csv'), ... generate_file_list('r', 101:112, 'customers.csv'), ... generate_file_list('rc', 101:108, 'customers.csv'), ... generate_file_list('c', 201:208, 'customers.csv'), ... generate_file_list('r', 201:211, 'customers.csv'), ... generate_file_list('rc', 201:208, 'customers.csv')]; distanceMatrix_files_list = [generate_file_list('c', 101:109, 'distanceMatrix.csv'), ... generate_file_list('r', 101:112, 'distanceMatrix.csv'), ... generate_file_list('rc', 101:108, 'distanceMatrix.csv'), ... generate_file_list('c', 201:208, 'distanceMatrix.csv'), ... generate_file_list('r', 201:211, 'distanceMatrix.csv'), ... generate_file_list('rc', 201:208, 'distanceMatrix.csv')]; initialPop_files_list = [generate_file_list('c', 101:109, 'initialPop.mat'), ... generate_file_list('r', 101:112, 'initialPop.mat'), ... generate_file_list('rc', 101:108, 'initialPop.mat'), ... generate_file_list('c', 201:208, 'initialPop.mat'), ... generate_file_list('r', 201:211, 'initialPop.mat'), ... generate_file_list('rc', 201:208, 'initialPop.mat')]; initialDistance_files_list = [generate_file_list('c', 101:109, 'initialDistance.mat'), ... generate_file_list('r', 101:112, 'initialDistance.mat'), ... generate_file_list('rc', 101:108, 'initialDistance.mat'), ... generate_file_list('c', 201:208, 'initialDistance.mat'), ... generate_file_list('r', 201:211, 'initialDistance.mat'), ... generate_file_list('rc', 201:208, 'initialDistance.mat')]; initialResults_files_list = [generate_file_list('c', 101:109, 'initialResults.csv'), ... generate_file_list('r', 101:112, 'initialResults.csv'), ... generate_file_list('rc', 101:108, 'initialResults.csv'), ... generate_file_list('c', 201:208, 'initialResults.csv'), ... generate_file_list('r', 201:211, 'initialResults.csv'), ... generate_file_list('rc', 201:208, 'initialResults.csv')]; % Experiment variables num_experiments = 30; num_cores = feature('numcores'); % MATLAB's equivalent to multiprocessing.cpu_count() % Global variables N = 300; % population size greedy_percent = 0.10; % percentage to be greedily initialized random_percent = 0.90; % percentage to be randomly initialized alpha = 100; beta = 0.001; % Loop through each dataset for i = 1:length(dataset_list) dataSet = dataset_list{i}; % File paths customer_file = fullfile(customer_path, customers_files_list{i}); dataInfo_file = fullfile(dataInfo_path, dataInfo_files_list{i}); distanceMatrix_file = fullfile(distanceMatrix_path, distanceMatrix_files_list{i}); % Read data df_customers = readtable(customer_file); df_distance_matrix = readtable(distanceMatrix_file); df_dataInfo = readtable(dataInfo_file); num_customers = height(df_customers); num_customers_depot = num_customers + 1; % Adjust customer indexing df_customers.Properties.RowNames = arrayfun(@num2str, 1:num_customers_depot, 'UniformOutput', false)'; number_random_chromosomes = floor(N * random_percent) - 1; number_nearest_nn_chromosomes = 1; number_greedy_chromosomes = N - number_random_chromosomes - 1; % int(N*greedy_percent) % Convert to arrays arr_distance_matrix = table2array(df_distance_matrix); % Calculate dataset radius dataset_radius = (max(arr_distance_matrix(:), [], 'omitnan') - min(arr_distance_matrix(:), [], 'omitnan')) / 2; % Customer array arr_customers = zeros(num_customers, 6); arr_customers(:,1) = (1:num_customers)'; arr_customers(:,2) = df_customers.demand; arr_customers(:,3) = df_customers.serviceTime; arr_customers(:,4) = df_customers.readyTime; arr_customers(:,5) = df_customers.dueTime; arr_customers(:,6) = df_customers.completeTime; % Fleet max details total_time = df_dataInfo.fleet_max_working_time(1); total_capacity = df_dataInfo.fleet_capacity(1); all_customers_list = 1:num_customers_depot - 1; % Loop through each experiment parfor experiment = 0:num_experiments-1 % Define experiment-specific output paths exp_initial_path = fullfile(initial_path, dataSet, sprintf('experiment%d', experiment)); if ~exist(exp_initial_path, 'dir') mkdir(exp_initial_path); end initial_pop_output_file = fullfile(exp_initial_path, initialPop_files_list{i}); initial_pop_distance_output_file = fullfile(exp_initial_path, initialDistance_files_list{i}); initial_result_output_file = fullfile(exp_initial_path, initialResults_files_list{i}); % Randomize radius radius = dataset_radius * (0.5 + (1-0.5) * rand()); % random.uniform(0.5,1) % Phase 1: Initial Population [pop_chromosome_routeList_array, pop_chromosome_distanceList_array, pop_num_routes, pop_total_distance] = ... initial_population(N, num_cores, number_random_chromosomes, number_greedy_chromosomes, ... number_nearest_nn_chromosomes, num_customers, arr_customers, arr_distance_matrix, ... total_capacity, total_time, all_customers_list, radius); % Phase 2: Initial Routing [pop_chromosome_routeList_array, pop_chromosome_distanceList_array, arr_pop_results] = ... initial_routing_phase2(num_cores, N, alpha, beta, pop_chromosome_routeList_array, ... pop_chromosome_distanceList_array, arr_distance_matrix, ... arr_customers, total_time, total_capacity); % Save outputs save(initial_pop_output_file, 'pop_chromosome_routeList_array'); save(initial_pop_distance_output_file, 'pop_chromosome_distanceList_array'); % Prepare initial results table df_initial_results = table(pop_num_routes, pop_total_distance, ... alpha * pop_num_routes + beta * pop_total_distance, ... 'VariableNames', {'num_vehicles', 'distance', 'fitness'}); writetable(df_initial_results, initial_result_output_file); fprintf('Dataset: %s, Experiment: %d completed.\n', dataSet, experiment); end fprintf('Dataset: %s processing completed.\n', dataSet); end2. Helper Functions
Each helper function from the Python scripts is translated into a separate MATLAB function. Below are the translations:
a.
initial_population.m% initial_population.m function [pop_chromosome_routeList_array, pop_chromosome_distanceList_array, pop_num_routes, pop_total_distance] = ... initial_population(N, num_cores, number_random_chromosomes, number_greedy_chromosomes, number_nearest_nn_chromosomes, ... num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, ... all_customers_list, radius) % Initialize cell arrays to store results chromosome_route_list = cell(N, 1); chromosome_distance_list = zeros(N, 1); pop_num_routes = zeros(N,1); pop_total_distance = zeros(N,1); % Initialize parallel pool if not already open pool = gcp('nocreate'); if isempty(pool) parpool('local', num_cores); end % Phase 1: Generate random and greedy chromosomes temp = cell(number_random_chromosomes + number_greedy_chromosomes, 1); parfor idx = 1:(number_random_chromosomes + number_greedy_chromosomes) if idx <= number_random_chromosomes [route, distance, num_routes, total_dist] = random_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list); else [route, distance, num_routes, total_dist] = greedy_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list, radius); end temp{idx} = {route, distance, num_routes, total_dist}; end % Phase 1 Results for idx = 1:length(temp) pop_chromosome_routeList_array{idx+1} = temp{idx}{1}; % Start from 2 since 1 is for nearest_neighbour pop_chromosome_distanceList_array(idx+1) = temp{idx}{2}; pop_num_routes(idx+1) = temp{idx}{3}; pop_total_distance(idx+1) = temp{idx}{4}; end % Phase 1: Nearest Neighbour Chromosome [nn_route, nn_distance, nn_num_routes, nn_total_distance] = nearest_neighbour(num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list); pop_chromosome_routeList_array{1} = nn_route; pop_chromosome_distanceList_array(1) = nn_distance; pop_num_routes(1) = nn_num_routes; pop_total_distance(1) = nn_total_distance; endb.
initial_routing_phase2.m% initial_routing_phase2.m function [pop_chromosome_routeList_array, pop_chromosome_distanceList_array, arr_pop_results] = ... initial_routing_phase2(num_cores, N, alpha, beta, pop_chromosome_routeList_array, pop_chromosome_distanceList_array, ... arr_distance_matrix, arr_customers, total_time, total_capacity) % Initialize cell arrays for updated routes and distances updated_routes = cell(N,1); updated_distances = zeros(N,1); arr_pop_results = zeros(N,3); % num_vehicles, distance, fitness % Initialize parallel pool if not already open pool = gcp('nocreate'); if isempty(pool) parpool('local', num_cores); end parfor i = 1:N [updated_route, updated_distance, num_vehicles, total_distance, fitness] = ... routing_phase2(alpha, beta, pop_chromosome_routeList_array{i}, pop_chromosome_distanceList_array(i), ... total_capacity, total_time, arr_distance_matrix, arr_customers); updated_routes{i} = updated_route; updated_distances(i) = updated_distance; arr_pop_results(i, :) = [num_vehicles, total_distance, fitness]; end % Assign updated values back pop_chromosome_routeList_array = updated_routes; pop_chromosome_distanceList_array = updated_distances; endc.
routing_phase2.m% routing_phase2.m function [chromosome, chromosome_distance_list, num_vehicles, total_distance, fitness] = ... routing_phase2(alpha, beta, chromosome, chromosome_distance_list, total_capacity, total_time, arr_distance_matrix, arr_customers) index_list = 1:(length(chromosome)+1); index_list(end) = 1; % MATLAB indices start at 1; replacing 0 with 1 for depot for i = 1:(length(index_list)-1) r1_index = index_list(i); r2_index = index_list(i+1); route_r1 = chromosome{r1_index}; route_r2 = chromosome{r2_index}; dist_r1 = chromosome_distance_list(r1_index); dist_r2 = chromosome_distance_list(r2_index); customer_r1 = route_r1(end-1); % Last customer in R1 % Insert customer_r1 into route_r2 at position 2 (after depot) route_r2 = [route_r2(1), customer_r1, route_r2(2:end)]; % Check validity [route_valid_status, new_r2_dist] = route_validity(route_r2, arr_distance_matrix, arr_customers, total_time, total_capacity); if route_valid_status if length(route_r1) > 3 new_r1_dist = dist_r1 - arr_distance_matrix(route_r1(end-1), 1) + arr_distance_matrix(route_r1(end-2), 1); if (dist_r1 + dist_r2) > (new_r1_dist + new_r2_dist) route_r1(end) = []; % Remove last depot chromosome{r1_index} = route_r1; chromosome{r2_index} = route_r2; chromosome_distance_list(r1_index) = new_r1_dist; chromosome_distance_list(r2_index) = new_r2_dist; end else % Remove route_r1 and update route_r2 chromosome(r1_index) = []; chromosome{r2_index} = route_r2; chromosome_distance_list(r2_index) = new_r2_dist; end end end % Calculate total distance and number of vehicles total_distance = sum(chromosome_distance_list); num_vehicles = length(chromosome); fitness = (alpha * num_vehicles) + (beta * total_distance); endd.
random_neighbour.m% random_neighbour.m function [chromosome_route_list, chromosome_distance_list, num_routes, total_distance] = random_neighbour( ... num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list) available_customers_list = all_customers_list; chromosome_route_list = {}; chromosome_distance_list = []; route_list = 0; % Depot represented by 0 curr_capacity = 0; curr_time = 0; curr_customer = 0; curr_distance = 0; while ~isempty(available_customers_list) arr_feasible_neighbours = feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, ... available_customers_list, total_time, total_capacity, ... curr_time, curr_capacity, curr_customer); if size(arr_feasible_neighbours,1) ~=0 feasible_neighbours_list = arr_feasible_neighbours(:,1)'; chosen_idx = randi(length(feasible_neighbours_list)); curr_customer = feasible_neighbours_list(chosen_idx); route_list(end+1) = curr_customer; curr_distance = curr_distance + arr_feasible_neighbours(chosen_idx,6); curr_capacity = arr_feasible_neighbours(chosen_idx,2); curr_time = arr_feasible_neighbours(chosen_idx,10); available_customers_list(available_customers_list == curr_customer) = []; else % End current route route_list(end+1) = 0; % Return to depot chromosome_route_list{end+1} = route_list; curr_distance = curr_distance + arr_distance_matrix(curr_customer+1,1); % MATLAB 1-based chromosome_distance_list(end+1) = curr_distance; % Initialize new route route_list = [0]; curr_capacity = 0; curr_time = 0; curr_customer = 0; curr_distance = 0; end end % End last route if route_list(end) ~=0 route_list(end+1) = 0; chromosome_route_list{end+1} = route_list; curr_distance = curr_distance + arr_distance_matrix(curr_customer+1,1); % MATLAB 1-based chromosome_distance_list(end+1) = curr_distance; end num_routes = length(chromosome_distance_list); total_distance = sum(chromosome_distance_list); ende.
greedy_neighbour.m% greedy_neighbour.m function [chromosome_route_list, chromosome_distance_list, num_routes, total_distance] = greedy_neighbour( ... num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list, radius) available_customers_list = all_customers_list; chromosome_route_list = {}; chromosome_distance_list = []; route_list = 0; % Depot curr_capacity = 0; curr_time = 0; curr_customer = 0; curr_distance = 0; while ~isempty(available_customers_list) arr_feasible_neighbours = feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, ... available_customers_list, total_time, total_capacity, ... curr_time, curr_capacity, curr_customer); if size(arr_feasible_neighbours,1) ~=0 % Filter by radius within_radius = arr_feasible_neighbours(arr_feasible_neighbours(:,6) <= radius, :); if ~isempty(within_radius) arr_feasible_neighbours = within_radius; end feasible_neighbours_list = arr_feasible_neighbours(:,1)'; chosen_idx = randi(length(feasible_neighbours_list)); curr_customer = feasible_neighbours_list(chosen_idx); route_list(end+1) = curr_customer; curr_distance = curr_distance + arr_feasible_neighbours(chosen_idx,6); curr_capacity = arr_feasible_neighbours(chosen_idx,2); curr_time = arr_feasible_neighbours(chosen_idx,10); available_customers_list(available_customers_list == curr_customer) = []; else % End current route route_list(end+1) = 0; % Return to depot chromosome_route_list{end+1} = route_list; curr_distance = curr_distance + arr_distance_matrix(curr_customer+1,1); % MATLAB 1-based chromosome_distance_list(end+1) = curr_distance; % Initialize new route route_list = [0]; curr_capacity = 0; curr_time = 0; curr_customer = 0; curr_distance = 0; end end % End last route if route_list(end) ~=0 route_list(end+1) = 0; chromosome_route_list{end+1} = route_list; curr_distance = curr_distance + arr_distance_matrix(curr_customer+1,1); % MATLAB 1-based chromosome_distance_list(end+1) = curr_distance; end num_routes = length(chromosome_distance_list); total_distance = sum(chromosome_distance_list); endf.
nearest_neighbour.m% nearest_neighbour.m function [chromosome_route_list, chromosome_distance_list, num_routes, total_distance] = nearest_neighbour( ... num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list) available_customers_list = all_customers_list; chromosome_route_list = {}; chromosome_distance_list = []; route_list = 0; % Depot curr_capacity = 0; curr_time = 0; curr_customer = 0; curr_distance = 0; while ~isempty(available_customers_list) arr_feasible_neighbours = feasible_neighbours(num_customers, arr_customers, arr_distance_matrix, ... available_customers_list, total_time, total_capacity, ... curr_time, curr_capacity, curr_customer); if size(arr_feasible_neighbours,1) ~=0 % Find minimum travel distance [min_dist, min_idx] = min(arr_feasible_neighbours(:,6)); curr_customer = arr_feasible_neighbours(min_idx,1); route_list(end+1) = curr_customer; curr_distance = curr_distance + min_dist; curr_capacity = arr_feasible_neighbours(min_idx,2); curr_time = arr_feasible_neighbours(min_idx,10); available_customers_list(available_customers_list == curr_customer) = []; else % End current route route_list(end+1) = 0; % Return to depot chromosome_route_list{end+1} = route_list; curr_distance = curr_distance + arr_distance_matrix(curr_customer+1,1); % MATLAB 1-based chromosome_distance_list(end+1) = curr_distance; % Initialize new route route_list = [0]; curr_capacity = 0; curr_time = 0; curr_customer = 0; curr_distance = 0; end end % End last route if route_list(end) ~=0 route_list(end+1) = 0; chromosome_route_list{end+1} = route_list; curr_distance = curr_distance + arr_distance_matrix(curr_customer+1,1); % MATLAB 1-based chromosome_distance_list(end+1) = curr_distance; end num_routes = length(chromosome_distance_list); total_distance = sum(chromosome_distance_list); endg.
feasible_neighbours.m% feasible_neighbours.m function arr_feasible_customers = feasible_neighbours( ... num_customers, arr_customers, arr_distance_matrix, available_customers_list, ... total_time, total_capacity, curr_time, curr_capacity, curr_customer) num_available = length(available_customers_list); arr_feasible_customers = zeros(num_available, 12); % Adjust indices for MATLAB (assuming customers are 1-indexed) available_indices = available_customers_list; % Populate arr_feasible_customers arr_feasible_customers(:,1:6) = arr_customers(available_indices, :); % Update capacities arr_feasible_customers(:,2) = arr_feasible_customers(:,2) + curr_capacity; % Distance from current customer to next arr_feasible_customers(:,6) = arr_distance_matrix(curr_customer+1, available_customers_list)'; % Arrival time arr_feasible_customers(:,7) = arr_feasible_customers(:,6) + curr_time; % Waiting time arr_feasible_customers(:,8) = arr_customers(available_indices,4) - arr_feasible_customers(:,7); arr_feasible_customers(:,8) = max(arr_feasible_customers(:,8), 0); % Start time arr_feasible_customers(:,9) = arr_feasible_customers(:,7) + arr_feasible_customers(:,8); % Finish time arr_feasible_customers(:,10) = arr_feasible_customers(:,9) + arr_feasible_customers(:,3); % Return time to depot arr_feasible_customers(:,11) = arr_distance_matrix(available_customers_list+1,1) + arr_feasible_customers(:,10); % Constraints: % Capacity capacity_ok = arr_feasible_customers(:,2) <= total_capacity; % Time to return to depot return_time_ok = arr_feasible_customers(:,11) <= total_time; % Start time within ready time start_time_ok = arr_feasible_customers(:,9) >= arr_feasible_customers(:,4); % Finish time within complete time finish_time_ok = arr_feasible_customers(:,10) <= arr_feasible_customers(:,6); % Apply constraints feasible_indices = capacity_ok & return_time_ok & start_time_ok & finish_time_ok; arr_feasible_customers = arr_feasible_customers(feasible_indices, :); endh.
route_validity.m% route_validity.m function [route_valid_status, curr_dist] = route_validity(route, arr_distance_matrix, arr_customers, total_time, total_capacity) route_valid_status = true; curr_dist = 0; % Initialize route parameters curr_dist = curr_dist + arr_distance_matrix(route(1)+1, route(2)+1); % MATLAB 1-based curr_capacity = arr_customers(route(2), 2); ready_time = arr_customers(route(2), 4); due_time = arr_customers(route(2), 5); curr_time = arr_distance_matrix(route(1)+1, route(2)+1); % Travel time waiting_time = max(0, ready_time - curr_time); curr_time = curr_time + waiting_time + arr_customers(route(2), 3); % Service time % Iterate through the route for i = 2:(length(route)-1) next_customer = route(i+1); travel_time = arr_distance_matrix(route(i)+1, next_customer+1); curr_dist = curr_dist + travel_time; curr_time = curr_time + travel_time; ready_time = arr_customers(next_customer, 4); due_time = arr_customers(next_customer, 5); service_time = arr_customers(next_customer, 3); curr_capacity = curr_capacity + arr_customers(next_customer, 2); if curr_time <= due_time && curr_capacity <= total_capacity waiting_time = max(0, ready_time - curr_time); if (curr_time + waiting_time >= ready_time) && (curr_time + waiting_time + service_time <= total_time) curr_time = curr_time + waiting_time + service_time; else route_valid_status = false; return; end else route_valid_status = false; return; end end % Final return to depot curr_dist = curr_dist + arr_distance_matrix(route(end-1)+1, route(end)+1); curr_time = curr_time + arr_distance_matrix(route(end-1)+1, route(end)+1); if curr_time > total_time route_valid_status = false; end end3. Notes and Considerations
-
Indexing Adjustments: MATLAB uses 1-based indexing, whereas Python uses 0-based indexing. In the translation:
- All customer indices are assumed to start at 1 in MATLAB.
- Depot is represented as
0in Python but adjusted appropriately in MATLAB (often by using1).
-
Parallel Processing: MATLAB's
parforloops are used to parallelize operations, similar to Python's multiprocessing. Ensure that MATLAB's Parallel Computing Toolbox is installed and configured. -
Data Storage:
- MATLAB does not have a direct equivalent to Python's
numpy.save. Instead,.matfiles are used to store data structures efficiently. - Adjustments are made to save and load data accordingly.
- MATLAB does not have a direct equivalent to Python's
-
Cell Arrays vs. MATLAB Arrays:
- Cell arrays are used to store variable-length routes (
chromosome_route_list) since MATLAB arrays require uniformity in dimensions. - Regular numeric arrays are used for fixed-size data like distances.
- Cell arrays are used to store variable-length routes (
-
Error Handling and Validation:
- Basic error checks are included to ensure route validity.
- Additional validation may be required based on specific data and requirements.
-
Function Dependencies: Ensure all helper functions are placed in the MATLAB path or the same directory as
main.m. -
Performance Considerations: MATLAB's performance characteristics differ from Python. Profiling and optimization may be necessary for large datasets or numerous experiments.
-
File Formats:
- The code assumes
.matand.csvfile formats. Ensure consistency with your data storage and usage.
- The code assumes
4. Running the MATLAB Code
-
Setup Directory Structure: Ensure that the directory structure matches the paths defined in the
main.mscript (solomon25_csv,initial, etc.). -
Place All Functions in Path: Save each helper function (
initial_population.m,initial_routing_phase2.m, etc.) in the same directory asmain.mor add their locations to MATLAB's search path. -
Execute Main Script: Run the
main.mscript in MATLAB. Ensure that the necessary data files are present in the specified directories. -
Monitor Outputs: The script will generate
.matand.csvfiles containing the initial populations, distances, and results for each experiment and dataset.
5. Example Function Call
Here's how you might call one of the helper functions independently for testing:
% Example: Testing the nearest_neighbour function num_customers = 5; arr_customers = [ 1, 10, 5, 0, 50, 55; 2, 15, 3, 10, 60, 63; 3, 20, 4, 20, 70, 74; 4, 25, 2, 30, 80, 82; 5, 30, 6, 40, 90, 96 ]; arr_distance_matrix = [ 0, 10, 20, 30, 40, 50; 10, 0, 15, 25, 35, 45; 20, 15, 0, 18, 28, 38; 30, 25, 18, 0, 12, 22; 40, 35, 28, 12, 0, 14; 50, 45, 38, 22, 14, 0 ]; total_capacity = 100; total_time = 100; all_customers_list = 1:5; [chromosome_route_list, chromosome_distance_list, num_routes, total_distance] = nearest_neighbour(... num_customers, arr_customers, arr_distance_matrix, total_capacity, total_time, all_customers_list); disp(chromosome_route_list); disp(chromosome_distance_list); disp(num_routes); disp(total_distance);6. Conclusion
Translating complex Python scripts to MATLAB requires meticulous attention to detail, especially when dealing with indexing and data structures. The provided translations aim to maintain the original logic and functionality while adapting to MATLAB's paradigms. Thorough testing is recommended to ensure that all functionalities perform as expected in the MATLAB environment.
If you encounter specific issues or require further assistance with particular sections of the code, feel free to ask!
- Main Script (