拟定了较大的输入数据
This commit is contained in:
parent
be23f5d29e
commit
6d4aa70df6
|
|
@ -17,7 +17,6 @@ class ChromosomeUtils:
|
|||
self.risk = risk_data # 风险企业数据
|
||||
self.supplier = supplier_data # 供应商数据
|
||||
self.I = order_data.I # 物料种类数
|
||||
|
||||
# 预计算每个物料的可选企业列表（0=风险企业，1+为供应商ID）
|
||||
self.material_optional_enterprises = self._get_material_optional_enterprises()
|
||||
# 每个物料的可选企业数量
|
||||
|
|
@ -85,7 +84,7 @@ class ChromosomeUtils:
|
|||
|
||||
def repair_capacity_layer(self, enterprise_layer: np.ndarray, capacity_layer: np.ndarray) -> np.ndarray:
|
||||
"""
|
||||
修复能力层：确保满足单物料产能约束和企业总产能约束
|
||||
修复能力层：满足单物料最小产能约束和企业总产能约束
|
||||
:param enterprise_layer: 企业层（用于确定哪些企业被选中）
|
||||
:param capacity_layer: 待修复的能力层
|
||||
:return: 修复后的能力层
|
||||
|
|
@ -94,7 +93,7 @@ class ChromosomeUtils:
|
|||
split_points = np.cumsum(self.material_enterprise_count)
|
||||
start = 0
|
||||
|
||||
# 1. 修复单物料产能约束（每个企业的单物料产能不超过上限）
|
||||
# 1. 修复单物料最小产能约束（每个企业的单物料产能不低于最小产能）
|
||||
for i in range(self.I):
|
||||
end = split_points[i]
|
||||
ents = self.material_optional_enterprises[i] # 可选企业列表
|
||||
|
|
@ -102,15 +101,16 @@ class ChromosomeUtils:
|
|||
enterprise_segment = enterprise_layer[start:end] # 企业选择状态
|
||||
for idx, ent in enumerate(ents):
|
||||
if enterprise_segment[idx] == 1: # 仅处理被选中的企业
|
||||
# 确定该企业的单物料最大产能
|
||||
# 确定该企业的单物料最小产能
|
||||
if ent == 0:
|
||||
max_cap = self.risk.C0_i_std[i] # 风险企业
|
||||
min_cap = self.risk.C0_i_min[i] # 风险企业最小产能
|
||||
else:
|
||||
supplier_id = ent - 1
|
||||
max_cap = self.supplier.Cj_i_std[supplier_id][i] # 供应商
|
||||
# 若产能不合法，重置为1到max_cap之间的随机值
|
||||
if segment[idx] <= 0 or segment[idx] > max_cap:
|
||||
segment[idx] = random.uniform(1, max_cap)
|
||||
min_cap = self.supplier.Cj_i_min[supplier_id][i] # 供应商最小产能
|
||||
|
||||
# 若产能低于最小产能，重置为最小产能到2倍最小产能之间的随机值（无上限约束，合理范围控制）
|
||||
if segment[idx] < min_cap:
|
||||
segment[idx] = random.uniform(min_cap, min_cap * 2)
|
||||
repaired[start:end] = segment
|
||||
start = end
|
||||
|
||||
|
|
@ -140,7 +140,7 @@ class ChromosomeUtils:
|
|||
max_total_cap = self.supplier.Cj_total_max[supplier_id] # 供应商总产能上限
|
||||
|
||||
if total_cap > max_total_cap: # 超出上限时缩放
|
||||
scale = max_total_cap / total_cap # 缩放比例（会出现浮点）
|
||||
scale = max_total_cap / total_cap # 缩放比例
|
||||
start = 0
|
||||
# 重新遍历，按比例缩小该企业所有物料的产能
|
||||
for i in range(self.I):
|
||||
|
|
@ -150,7 +150,14 @@ class ChromosomeUtils:
|
|||
capacity_segment = repaired[start:end]
|
||||
for idx, e in enumerate(ents):
|
||||
if e == ent and enterprise_segment[idx] == 1:
|
||||
capacity_segment[idx] *= scale # 按比例缩小
|
||||
# 缩放后仍需保证不低于最小产能
|
||||
scaled_cap = capacity_segment[idx] * scale
|
||||
if ent == 0:
|
||||
min_cap = self.risk.C0_i_min[i]
|
||||
else:
|
||||
supplier_id = e - 1
|
||||
min_cap = self.supplier.Cj_i_min[supplier_id][i]
|
||||
capacity_segment[idx] = max(scaled_cap, min_cap)
|
||||
repaired[start:end] = capacity_segment
|
||||
start = end
|
||||
|
||||
|
|
|
|||
|
|
@ -1,36 +1,28 @@
|
|||
# 数据结构定义：存储订单、企业、供应商数据及算法配置
|
||||
|
||||
|
||||
class OrderData:
|
||||
"""订单数据类：存储物料需求、交货期、成本等信息"""
|
||||
|
||||
def __init__(self):
|
||||
self.I = 5 # 物料种类数
|
||||
self.Q = [250, 300, 200, 350, 280] # 各物料的需求数量
|
||||
self.Dd = 12 # 需求交货期（单位：时间）
|
||||
self.P0 = [50, 80, 60, 70, 90] # 风险企业的单位采购价
|
||||
self.Q = [6000, 12000, 20000, 7500, 13500] # 各物料的需求数量
|
||||
self.Dd = 30 # 需求交货期（单位：时间）
|
||||
self.P0 = [45, 30, 30, 50, 40] # 风险企业的单位采购价
|
||||
self.T0 = [5, 8, 6, 7, 9] # 风险企业的单位运输成本
|
||||
self.transport_speed = 10 # 运输速度（单位：距离/时间）
|
||||
|
||||
|
||||
class RiskEnterpriseData:
|
||||
"""风险企业数据类：存储风险企业的产能、距离等信息"""
|
||||
|
||||
def __init__(self):
|
||||
self.I = 5 # 物料种类数（与订单一致）
|
||||
self.C0_i_std = [40, 50, 35, 45, 48] # 单物料的单位时间标准产能
|
||||
self.C0_total_max = 100 # 总产能上限（单位时间）
|
||||
self.distance = 10 # 与需求点的距离
|
||||
|
||||
self.C0_i_min = [50, 100, 150, 80, 100] # 单物料的单位时间最小产能
|
||||
self.C0_total_max = 900 # 总产能上限（单位时间）
|
||||
self.distance = 20 # 与需求点的距离
|
||||
|
||||
class SupplierData:
|
||||
"""供应商数据类：存储各供应商的产能、价格、距离等信息"""
|
||||
|
||||
def __init__(self, I=5):
|
||||
self.I = I # 物料种类数
|
||||
self.supplier_count = 4 # 供应商数量
|
||||
self.names = ["S0", "S1", "S2", "S3"] # 供应商名称
|
||||
|
||||
# 能否生产某物料的矩阵（supplier_count × I），1=能生产，0=不能
|
||||
self.can_produce = [
|
||||
[1, 1, 1, 1, 1],
|
||||
|
|
@ -38,76 +30,63 @@ class SupplierData:
|
|||
[0, 1, 0, 1, 0],
|
||||
[0, 0, 1, 1, 1]
|
||||
]
|
||||
|
||||
# 单物料单位时间标准产能（supplier_count × I）
|
||||
self.Cj_i_std = [
|
||||
[20, 18, 15, 22, 25],
|
||||
[25, 0, 30, 0, 28],
|
||||
[0, 22, 0, 35, 0],
|
||||
[0, 0, 20, 30, 22]
|
||||
# 单物料单位时间最小产能（supplier_count × I），0表示不能生产该物料
|
||||
self.Cj_i_min = [
|
||||
[30, 80, 100, 60, 80],
|
||||
[60, 0, 180, 0, 120],
|
||||
[0, 150, 0, 120, 0],
|
||||
[0, 0, 170, 105, 115]
|
||||
]
|
||||
|
||||
# 供应商单位时间的最大总产能（supplier_count）
|
||||
self.Cj_total_max = [120, 110, 100, 95]
|
||||
|
||||
self.Cj_total_max = [700, 800, 600, 850]
|
||||
# 最小起订量（supplier_count × I）
|
||||
self.MinOrder = [
|
||||
[20, 20, 15, 25, 20],
|
||||
[30, 0, 25, 0, 30],
|
||||
[0, 25, 0, 30, 0],
|
||||
[0, 0, 20, 35, 25]
|
||||
[800, 1500, 3000, 800, 1500],
|
||||
[1000, 0, 3500, 0, 1800],
|
||||
[0, 1700, 0, 1000, 0],
|
||||
[0, 0, 2500, 500, 1000]
|
||||
]
|
||||
|
||||
# 最大供应量（supplier_count × I）
|
||||
self.MaxOrder = [
|
||||
[100, 150, 80, 120, 130],
|
||||
[120, 0, 100, 0, 110],
|
||||
[0, 140, 0, 150, 0],
|
||||
[0, 0, 90, 130, 100]
|
||||
[5000, 10000, 18000, 6500, 11000],
|
||||
[8000, 0, 25000, 0, 15000],
|
||||
[0, 8000, 0, 6000, 0],
|
||||
[0, 0, 20000, 7500, 13500]
|
||||
]
|
||||
|
||||
# 单位采购价格（supplier_count × I）
|
||||
self.P_ij = [
|
||||
[60, 85, 70, 80, 100],
|
||||
[65, 0, 75, 0, 105],
|
||||
[0, 90, 0, 85, 0],
|
||||
[0, 0, 78, 88, 98]
|
||||
[50, 35, 28, 47, 38],
|
||||
[43, 0, 28, 0, 36],
|
||||
[0, 31, 0, 52, 0],
|
||||
[0, 0, 32, 52, 43]
|
||||
]
|
||||
|
||||
# 单位运输成本（supplier_count × I）
|
||||
self.T_ij = [
|
||||
[7, 9, 8, 10, 12],
|
||||
[6, 0, 9, 0, 11],
|
||||
[0, 10, 0, 12, 0],
|
||||
[0, 0, 10, 13, 14]
|
||||
[6, 9, 8, 9, 12],
|
||||
[4, 0, 5, 0, 15],
|
||||
[0, 10, 0, 7, 0],
|
||||
[0, 0, 8, 9, 11]
|
||||
]
|
||||
|
||||
# 供应商与需求点的距离（supplier_count）
|
||||
self.distance = [45, 35, 60, 50]
|
||||
|
||||
self.distance = [60, 50, 70, 40]
|
||||
|
||||
class Config:
|
||||
"""算法参数配置类：存储NSGA-II的各类参数"""
|
||||
|
||||
def __init__(self):
|
||||
# 种群参数
|
||||
self.pop_size = 300 # 种群大小
|
||||
self.pop_size = 100 # 种群大小
|
||||
self.N1_ratio = 0.2 # 优先成本的种群比例
|
||||
self.N2_ratio = 0.2 # 优先延期的种群比例
|
||||
self.N3_ratio = 0.3 # 强制风险企业的种群比例
|
||||
self.N4_ratio = 0.3 # 随机种群比例
|
||||
|
||||
# 遗传操作参数
|
||||
self.crossover_prob = 0.8 # 交叉概率
|
||||
self.mutation_prob = 0.2 # 变异概率
|
||||
self.max_generations = 800 # 最大进化代数
|
||||
|
||||
self.mutation_prob = 0.3 # 变异概率
|
||||
self.max_generations = 300 # 最大进化代数
|
||||
# 惩罚系数
|
||||
self.delta = 1.3 # 变更惩罚系数
|
||||
self.gamma = 500 # 提前交付惩罚系数
|
||||
|
||||
# 早停参数
|
||||
self.early_stop_patience = 50 # 连续多少代无改进则早停
|
||||
|
||||
# 目标函数数量
|
||||
self.objective_num = 2 # 双目标（成本+延期）
|
||||
27
encoder.py
27
encoder.py
|
|
@ -5,10 +5,8 @@ from chromosome_utils import ChromosomeUtils
|
|||
from objective_calculator import ObjectiveCalculator
|
||||
from nsga2 import NSGA2
|
||||
|
||||
|
||||
class Encoder:
|
||||
"""种群初始化编码器：生成初始种群，包含多种初始化策略"""
|
||||
|
||||
def __init__(self, config: Config, utils: ChromosomeUtils):
|
||||
"""
|
||||
初始化编码器
|
||||
|
|
@ -18,7 +16,6 @@ class Encoder:
|
|||
self.config = config
|
||||
self.utils = utils
|
||||
self.pop_size = config.pop_size # 种群总大小
|
||||
|
||||
# 按比例分配四种初始化策略的个体数量
|
||||
self.N1 = int(config.N1_ratio * self.pop_size) # 优先最小化成本的个体数
|
||||
self.N2 = int(config.N2_ratio * self.pop_size) # 优先最小化延期的个体数
|
||||
|
|
@ -34,11 +31,9 @@ class Encoder:
|
|||
enterprise_layer = [] # 企业选择层（0/1）
|
||||
capacity_layer = [] # 产能层
|
||||
quantity_layer = [] # 数量层
|
||||
|
||||
for i in range(self.utils.I): # 遍历每种物料
|
||||
ent_count = self.utils.material_enterprise_count[i] # 当前物料的可选企业数量
|
||||
ents = self.utils.material_optional_enterprises[i] # 可选企业列表（0为风险企业）
|
||||
|
||||
# 1. 生成企业层（随机选择至少1个企业）
|
||||
e_genes = np.zeros(ent_count) # 初始化企业选择为0（未选择）
|
||||
select_count = random.randint(1, ent_count) # 随机选择1到ent_count个企业
|
||||
|
|
@ -47,19 +42,18 @@ class Encoder:
|
|||
if force_risk_enterprise:
|
||||
e_genes[0] = 1 # 强制选择风险企业（索引0）
|
||||
enterprise_layer.extend(e_genes)
|
||||
|
||||
# 2. 生成能力层（为选中的企业随机分配产能）
|
||||
# 2. 生成能力层（为选中的企业分配不低于最小产能的随机产能）
|
||||
c_genes = np.zeros(ent_count)
|
||||
for idx, ent in enumerate(ents):
|
||||
if e_genes[idx] == 1: # 仅为选中的企业分配产能
|
||||
if ent == 0:
|
||||
max_cap = self.utils.risk.C0_i_std[i] # 风险企业的最大产能
|
||||
min_cap = self.utils.risk.C0_i_min[i] # 风险企业的最小产能
|
||||
else:
|
||||
supplier_id = ent - 1
|
||||
max_cap = self.utils.supplier.Cj_i_std[supplier_id][i] # 供应商的最大产能
|
||||
c_genes[idx] = random.uniform(1, max_cap) # 随机产能
|
||||
min_cap = self.utils.supplier.Cj_i_min[supplier_id][i] # 供应商的最小产能
|
||||
# 生成不低于最小产能的随机值（无上限，用2倍最小产能控制初始范围）
|
||||
c_genes[idx] = random.uniform(min_cap, min_cap * 2)
|
||||
capacity_layer.extend(c_genes)
|
||||
|
||||
# 3. 生成数量层（为选中的企业随机分配数量）
|
||||
q_genes = np.zeros(ent_count)
|
||||
for idx, ent in enumerate(ents):
|
||||
|
|
@ -71,7 +65,6 @@ class Encoder:
|
|||
max_q = self.utils.supplier.MaxOrder[supplier_id][i] # 供应商的最大供应量
|
||||
q_genes[idx] = random.uniform(1, max_q) # 随机数量
|
||||
quantity_layer.extend(q_genes)
|
||||
|
||||
# 合并三层并修复染色体（确保满足所有约束）
|
||||
chromosome = self.utils._merge_chromosome(
|
||||
np.array(enterprise_layer), np.array(capacity_layer), np.array(quantity_layer)
|
||||
|
|
@ -88,12 +81,10 @@ class Encoder:
|
|||
pop2 = self._initialize_by_objective(self.N2, "tardiness") # 优先延期
|
||||
pop3 = self._initialize_by_risk_enterprise(self.N3) # 强制风险企业
|
||||
pop4 = self._initialize_random(self.N4) # 随机生成
|
||||
|
||||
# 过滤空数组（避免合并时报错）
|
||||
population_list = [p for p in [pop1, pop2, pop3, pop4] if len(p) > 0]
|
||||
if len(population_list) == 0: # 所有子种群都为空时返回空数组
|
||||
return np.array([])
|
||||
|
||||
# 合并子种群并打乱顺序
|
||||
population = np.vstack(population_list)
|
||||
np.random.shuffle(population)
|
||||
|
|
@ -108,16 +99,13 @@ class Encoder:
|
|||
"""
|
||||
if count <= 0: # 数量为0时返回空数组
|
||||
return np.array([])
|
||||
|
||||
# 生成候选解（数量为count的3倍或至少20个，确保有足够选择空间）
|
||||
candidate_count = max(3 * count, 20)
|
||||
candidates = [self._generate_random_chromosome() for _ in range(candidate_count)]
|
||||
|
||||
# 计算候选解的目标函数值
|
||||
calculator = ObjectiveCalculator(self.utils.order, self.utils.risk, self.utils.supplier, self.utils,
|
||||
self.config)
|
||||
objectives = [calculator.calculate_objectives(chrom) for chrom in candidates]
|
||||
|
||||
# 按目标函数排序并选择前count个
|
||||
if objective_type == "cost":
|
||||
# 按成本升序排序（成本越小越优）
|
||||
|
|
@ -125,7 +113,6 @@ class Encoder:
|
|||
else:
|
||||
# 按延期升序排序（延期越小越优）
|
||||
sorted_indices = sorted(range(candidate_count), key=lambda x: objectives[x][1])
|
||||
|
||||
return np.array([candidates[i] for i in sorted_indices[:count]])
|
||||
|
||||
def _initialize_by_risk_enterprise(self, count: int) -> np.ndarray:
|
||||
|
|
@ -136,18 +123,15 @@ class Encoder:
|
|||
"""
|
||||
if count <= 0:
|
||||
return np.array([])
|
||||
|
||||
# 生成强制选择风险企业的候选解
|
||||
candidate_count = max(2 * count, 20)
|
||||
candidates = [self._generate_random_chromosome(force_risk_enterprise=True) for _ in range(candidate_count)]
|
||||
|
||||
# 计算目标函数并进行非支配排序
|
||||
calculator = ObjectiveCalculator(self.utils.order, self.utils.risk, self.utils.supplier, self.utils,
|
||||
self.config)
|
||||
objectives = [calculator.calculate_objectives(chrom) for chrom in candidates]
|
||||
nsga2 = NSGA2(candidate_count, 2) # 2个目标函数
|
||||
ranks, fronts = nsga2.fast_non_dominated_sort(objectives) # 非支配排序
|
||||
|
||||
# 从帕累托前沿开始选择，直到满足数量
|
||||
selected = []
|
||||
for front in fronts:
|
||||
|
|
@ -157,7 +141,6 @@ class Encoder:
|
|||
# 前沿数量超过剩余需求时，取部分
|
||||
selected.extend([candidates[i] for i in front[:count - len(selected)]])
|
||||
break
|
||||
|
||||
return np.array(selected)
|
||||
|
||||
def _initialize_random(self, count: int) -> np.ndarray:
|
||||
|
|
|
|||
|
|
@ -3,10 +3,8 @@ import numpy as np
|
|||
from data_structures import Config # 算法配置参数类
|
||||
from chromosome_utils import ChromosomeUtils # 染色体工具类
|
||||
|
||||
|
||||
class GeneticOperator:
|
||||
"""遗传操作类：实现交叉和变异操作，用于产生新个体"""
|
||||
|
||||
def __init__(self, config: Config, utils: ChromosomeUtils):
|
||||
"""
|
||||
初始化遗传操作器
|
||||
|
|
@ -26,17 +24,14 @@ class GeneticOperator:
|
|||
length = self.utils.chromosome_length # 染色体总长度
|
||||
if length < 2: # 染色体长度不足2时无法交叉，直接返回父代副本
|
||||
return parent1.copy(), parent2.copy()
|
||||
|
||||
# 随机选择两个交叉点（point1 < point2）
|
||||
point1 = random.randint(0, length // 2)
|
||||
point2 = random.randint(point1 + 1, length - 1)
|
||||
|
||||
# 复制父代基因并交换片段
|
||||
child1 = parent1.copy()
|
||||
child2 = parent2.copy()
|
||||
child1[point1:point2] = parent2[point1:point2] # 交换point1到point2之间的基因
|
||||
child2[point1:point2] = parent1[point1:point2]
|
||||
|
||||
# 修复染色体（确保满足约束条件）
|
||||
child1 = self.utils.repair_chromosome(child1)
|
||||
child2 = self.utils.repair_chromosome(child2)
|
||||
|
|
@ -49,14 +44,11 @@ class GeneticOperator:
|
|||
:return: 变异后的染色体（经过修复）
|
||||
"""
|
||||
mutated = chromosome.copy() # 复制原始染色体，避免直接修改
|
||||
|
||||
# 拆分染色体为三层
|
||||
enterprise_layer, capacity_layer, quantity_layer = self.utils._split_chromosome(mutated)
|
||||
split_points = np.cumsum(self.utils.material_enterprise_count)
|
||||
|
||||
# ========== 第一步：记录原始企业层状态 ==========
|
||||
original_enterprise_layer = enterprise_layer.copy()
|
||||
|
||||
# ========== 第二步：企业层变异 ==========
|
||||
start = 0
|
||||
for i in range(self.utils.I):
|
||||
|
|
@ -65,37 +57,30 @@ class GeneticOperator:
|
|||
if random.random() < self.config.mutation_prob:
|
||||
enterprise_layer[idx] = 1 - enterprise_layer[idx]
|
||||
start = end
|
||||
|
||||
# 修复企业层
|
||||
enterprise_layer = self.utils.repair_enterprise_layer(enterprise_layer)
|
||||
|
||||
# ========== 第三步：同步更新能力和数量层 ==========
|
||||
start = 0
|
||||
for i in range(self.utils.I):
|
||||
end = split_points[i]
|
||||
ents = self.utils.material_optional_enterprises[i]
|
||||
|
||||
for idx in range(start, end):
|
||||
current_ent_selected = enterprise_layer[idx] # 变异后的选择状态
|
||||
original_ent_selected = original_enterprise_layer[idx] # 变异前的选择状态
|
||||
|
||||
# 情况1：企业从选中变为未选中（1→0）
|
||||
if original_ent_selected == 1 and current_ent_selected == 0:
|
||||
capacity_layer[idx] = 0 # 产能设为0
|
||||
quantity_layer[idx] = 0 # 数量设为0
|
||||
|
||||
# 情况2：企业从未选中变为选中（0→1）
|
||||
elif original_ent_selected == 0 and current_ent_selected == 1:
|
||||
ent = ents[idx - start]
|
||||
|
||||
# 初始化产能
|
||||
# 初始化产能（不低于最小产能）
|
||||
if ent == 0:
|
||||
max_cap = self.utils.risk.C0_i_std[i]
|
||||
min_cap = self.utils.risk.C0_i_min[i]
|
||||
else:
|
||||
supplier_id = ent - 1
|
||||
max_cap = self.utils.supplier.Cj_i_std[supplier_id][i]
|
||||
capacity_layer[idx] = random.uniform(1, max_cap)
|
||||
|
||||
min_cap = self.utils.supplier.Cj_i_min[supplier_id][i]
|
||||
capacity_layer[idx] = random.uniform(min_cap, min_cap * 2)
|
||||
# 初始化数量
|
||||
qi = self.utils.order.Q[i]
|
||||
if ent == 0:
|
||||
|
|
@ -104,43 +89,36 @@ class GeneticOperator:
|
|||
supplier_id = ent - 1
|
||||
max_q = self.utils.supplier.MaxOrder[supplier_id][i]
|
||||
quantity_layer[idx] = random.uniform(1, max_q)
|
||||
|
||||
# 情况3：企业保持选中（1→1）或保持未选中（0→0）
|
||||
# 保持不变，后续的变异步骤会处理
|
||||
|
||||
start = end
|
||||
|
||||
# ========== 第四步：能力层变异（仅对保持选中的企业） ==========
|
||||
start = 0
|
||||
for i in range(self.utils.I):
|
||||
end = split_points[i]
|
||||
ents = self.utils.material_optional_enterprises[i]
|
||||
e_segment = enterprise_layer[start:end]
|
||||
|
||||
for idx in range(start, end):
|
||||
# 只对保持选中的企业进行变异（1→1的情况）
|
||||
if e_segment[idx - start] == 1 and original_enterprise_layer[idx] == 1:
|
||||
if random.random() < self.config.mutation_prob:
|
||||
ent = ents[idx - start]
|
||||
if ent == 0:
|
||||
max_cap = self.utils.risk.C0_i_std[i]
|
||||
min_cap = self.utils.risk.C0_i_min[i]
|
||||
else:
|
||||
supplier_id = ent - 1
|
||||
max_cap = self.utils.supplier.Cj_i_std[supplier_id][i]
|
||||
capacity_layer[idx] = random.uniform(1, max_cap)
|
||||
|
||||
min_cap = self.utils.supplier.Cj_i_min[supplier_id][i]
|
||||
# 变异后产能不低于最小产能
|
||||
capacity_layer[idx] = random.uniform(min_cap, min_cap * 2)
|
||||
start = end
|
||||
|
||||
# 修复能力层（考虑所有选中企业的产能）
|
||||
capacity_layer = self.utils.repair_capacity_layer(enterprise_layer, capacity_layer)
|
||||
|
||||
# ========== 第五步：数量层变异（仅对保持选中的企业） ==========
|
||||
start = 0
|
||||
for i in range(self.utils.I):
|
||||
end = split_points[i]
|
||||
ents = self.utils.material_optional_enterprises[i]
|
||||
e_segment = enterprise_layer[start:end]
|
||||
|
||||
for idx in range(start, end):
|
||||
# 只对保持选中的企业进行变异（1→1的情况）
|
||||
if e_segment[idx - start] == 1 and original_enterprise_layer[idx] == 1:
|
||||
|
|
@ -153,12 +131,9 @@ class GeneticOperator:
|
|||
supplier_id = ent - 1
|
||||
max_q = self.utils.supplier.MaxOrder[supplier_id][i]
|
||||
quantity_layer[idx] = random.uniform(1, max_q)
|
||||
|
||||
start = end
|
||||
|
||||
# 修复数量层（考虑所有选中企业的分配）
|
||||
quantity_layer = self.utils.repair_quantity_layer(enterprise_layer, quantity_layer)
|
||||
|
||||
# 合并三层
|
||||
mutated = self.utils._merge_chromosome(enterprise_layer, capacity_layer, quantity_layer)
|
||||
return mutated
|
||||
Loading…
Reference in New Issue