140 lines
5.6 KiB
Python
140 lines
5.6 KiB
Python
import numpy as np
|
||
import pandas as pd
|
||
import networkx as nx
|
||
from objective_func import objective_func
|
||
from convert_to_partition import convert_to_partition
|
||
import timeit
|
||
|
||
# x = np.array([8, 3, 6, 1, 3, 7, 2, 2, 2, 4, 6, 5, 8, 5, 5, 3, 1, 3, 2, 2, 7, 7,4, 8, 1, 7, 1, 1, 3, 7, 4, 7, 2, 6, 2, 8, 1, 3, 1, 6])
|
||
|
||
def liziqunfun(x,yanlinks,road_velocity,dadr,bus_line_sequence,result_array):
|
||
|
||
# Assuming x is a list of integers.
|
||
if isinstance(x, list):
|
||
x = x[0]
|
||
|
||
# Use numpy advanced (integer) indexing to avoid for loop
|
||
yanlinks[:, 11] = x[yanlinks[:, 10].astype(int) - 1]
|
||
|
||
data_path = r''
|
||
df2 = pd.read_csv(data_path + 'links_test1.csv')
|
||
df2['L'] = yanlinks[:, 11]
|
||
df1 = pd.read_csv(data_path + 'bus_road_sequence_processed.csv', header=None, encoding='gbk')
|
||
|
||
|
||
replace_dict = df2.set_index(df2.columns[5])[df2.columns[11]].to_dict()
|
||
df1.iloc[:, 1:] = df1.iloc[:, 1:].applymap(replace_dict.get)
|
||
|
||
bus_line_partition = df1.copy()
|
||
bus_line_partition = bus_line_partition.drop(bus_line_partition.columns[0], axis=1)
|
||
bus_line_partition = bus_line_partition.fillna(0)
|
||
bus_line_partition = bus_line_partition.astype(int)
|
||
bus_line_partition = bus_line_partition.to_numpy()
|
||
|
||
|
||
|
||
partition = convert_to_partition(yanlinks)
|
||
G = nx.Graph()
|
||
G.add_edges_from([(row[0], row[1], {'weight': 1}) for row in yanlinks])
|
||
|
||
# 删除全为零的行和列
|
||
rows_non_zero = np.any(dadr != 0, axis=1)
|
||
cols_non_zero = np.any(dadr != 0, axis=0)
|
||
adr = dadr[rows_non_zero][:, cols_non_zero]
|
||
# 初始化 badr
|
||
badr = np.zeros((len(np.unique(x)), len(np.unique(x))))
|
||
|
||
# 为了避免在循环中多次计算，提前计算好非零元素的索引
|
||
non_zero_indices = np.argwhere(adr)
|
||
|
||
for i, j in non_zero_indices:
|
||
if x[i] != x[j]:
|
||
badr[x[i] - 1, x[j] - 1] = 1
|
||
badr[x[j] - 1, x[i] - 1] = 1
|
||
|
||
y = {}
|
||
z = 0
|
||
max_bus_line_partition = np.max(bus_line_partition)
|
||
# 遍历 route_velocity 中的每个值
|
||
road_velocity_dict = {row[0]: row[1] for row in road_velocity}
|
||
|
||
def get_velocity(value):
|
||
return road_velocity_dict.get(value, 0)
|
||
vectorized_get_velocity = np.vectorize(get_velocity)
|
||
for i in range(1, max_bus_line_partition + 1):
|
||
wu = np.where(badr[:, i - 1] == 1)[0]
|
||
bus_line_partition_id = [index for index, row in zip(range(len(bus_line_partition)), bus_line_partition) if
|
||
i in row]
|
||
route_sequence = bus_line_sequence[bus_line_partition_id, :]
|
||
|
||
bus_line_route_velocity = vectorized_get_velocity(route_sequence)
|
||
|
||
# 1. 从 bus_line_partition 提取 rows
|
||
bus_line_partition_id = bus_line_partition[bus_line_partition_id]
|
||
|
||
# 2. 对每一行使用 numpy 的 unique 函数，进行去重操作
|
||
unique_matching_rows = [np.unique(row) for row in bus_line_partition_id]
|
||
|
||
# 3. 筛选与 wu 交集的元素
|
||
bus_region_direction = [np.intersect1d(row, wu) for row in unique_matching_rows]
|
||
|
||
# 4. 计算最大长度以进行填充操作
|
||
max_length = max(len(item) for item in bus_region_direction)
|
||
|
||
# 5. 使用0填充长度不足的部分，并将数组转换为二维数组
|
||
bus_region_direction = np.array(
|
||
[np.pad(item, (0, max_length - len(item)), 'constant', constant_values=0) for item in bus_region_direction])
|
||
|
||
# 6. 对 bus_region_direction 进行排序
|
||
sorted_bus_region_direction = np.sort(bus_region_direction, axis=1)
|
||
|
||
# 7. 获取唯一的 bus_region_direction
|
||
unique_sorted_bus_region_direction = np.unique(sorted_bus_region_direction, axis=0)
|
||
|
||
# 8. 初始化 same_direction_fitness 字典
|
||
same_direction_fitness = {}
|
||
|
||
# 计算整个数组的行匹配
|
||
matching_rows = np.all(sorted_bus_region_direction[:, None] == unique_sorted_bus_region_direction, axis=2)
|
||
|
||
# 计算总线路数
|
||
total_bus_lines = bus_line_route_velocity.shape[0]
|
||
|
||
total_bus_lines = len(matching_rows) # 这个应该在循环外部计算一次
|
||
|
||
for aa, row_match in enumerate(matching_rows.T):
|
||
matching_row_indices = np.where(row_match)[0]
|
||
same_direction_bus_line_route_velocity = bus_line_route_velocity[matching_row_indices]
|
||
same_direction_bus_line_partition_id = bus_line_partition_id[matching_row_indices]
|
||
|
||
line_proportion = len(matching_row_indices) / total_bus_lines
|
||
|
||
# 使用NumPy的向量化操作
|
||
matching_values = same_direction_bus_line_route_velocity[
|
||
np.where(same_direction_bus_line_partition_id == i)]
|
||
matching_values = matching_values[matching_values != 0]
|
||
|
||
# 使用numpy来决定是否计算方差
|
||
variance = np.var(matching_values) if len(matching_values) > 1 else 0
|
||
|
||
# 使用方差和线路比例更新 same_direction_fitness
|
||
same_direction_fitness[aa] = line_proportion * variance
|
||
|
||
links_number = np.count_nonzero(yanlinks[:,11]==i)-yanlinks.shape[0]/max_bus_line_partition
|
||
z +=abs(links_number)
|
||
y[i] = sum(same_direction_fitness.values())
|
||
# unique_values, unique_indices = np.unique(yanlinks[:, 11], return_index=True)
|
||
# Tvb = 0
|
||
# for i in unique_values:
|
||
# selected_values = yanlinks[yanlinks[:, 11] == i][:, 5]
|
||
# variance = np.var(selected_values)
|
||
# Tvb += variance
|
||
a = objective_func(G, partition,yanlinks,result_array)
|
||
y = sum(y.values())
|
||
# y = np.sqrt(y**2 + a**2 + z**2)
|
||
# y = np.concatenate((y,a,z),axis=1)
|
||
|
||
# print(f'适应度值为:{y}')
|
||
return y,a,z
|
||
# aa = liziqunfun(x)
|
||
# print(aa) |