HaoYizhi
/
IIabm
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

larger sample than population when replace is False

This commit is contained in:
HaoYizhi 2023-06-11 11:49:15 +08:00
parent 49d3c6791e
commit 5387268d48
3 changed files with 128 additions and 48 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
__pycache__/model.cpython-38.pyc
View File

Binary file not shown.

149
model.py
View File

@ -19,7 +19,7 @@ class Model(ap.Model):
self.dct_lst_remove_firm_prod = self.p.dct_lst_init_remove_firm_prod self.dct_lst_remove_firm_prod = self.p.dct_lst_init_remove_firm_prod
self.int_n_max_trial = int(self.p.n_max_trial) self.int_n_max_trial = int(self.p.n_max_trial)
self.int_netw_sply_prf_size = int(self.p.netw_sply_prf_n) self.int_netw_sply_prf_n = int(self.p.netw_sply_prf_n)
self.flt_netw_sply_prf_size = float(self.p.netw_sply_prf_size) self.flt_netw_sply_prf_size = float(self.p.netw_sply_prf_size)
self.str_cap_limit_prob_type = str(self.p.cap_limit_prob_type) self.str_cap_limit_prob_type = str(self.p.cap_limit_prob_type)
self.flt_cap_limit_level = float(self.p.cap_limit_level) self.flt_cap_limit_level = float(self.p.cap_limit_level)
@ -65,58 +65,111 @@ class Model(ap.Model):
firm_prod_labels_dict[code] = firm_prod.loc[code].to_dict() firm_prod_labels_dict[code] = firm_prod.loc[code].to_dict()
nx.set_node_attributes(G_FirmProd, firm_prod_labels_dict) nx.set_node_attributes(G_FirmProd, firm_prod_labels_dict)
# add edge to G_firm according to G_bom # # add edge to G_firm according to G_bom
for node in nx.nodes(G_Firm): # for node in nx.nodes(G_Firm):
lst_pred_product_code = [] # lst_pred_product_code = []
for product_code in G_Firm.nodes[node]['Product_Code']: # for product_code in G_Firm.nodes[node]['Product_Code']:
lst_pred_product_code += list(G_bom.predecessors(product_code)) # lst_pred_product_code += list(G_bom.predecessors(product_code))
lst_pred_product_code = list(set(lst_pred_product_code)) # lst_pred_product_code = list(set(lst_pred_product_code))
# to generate consistant graph # # to generate consistant graph
lst_pred_product_code = list(sorted(lst_pred_product_code)) # lst_pred_product_code = list(sorted(lst_pred_product_code))
for pred_product_code in lst_pred_product_code: # for pred_product_code in lst_pred_product_code:
# for each product predecessor (component) the firm need # # for each product predecessor (component) the firm need
# get a list of firm producing this component # # get a list of firm producing this component
lst_pred_firm = \ # lst_pred_firm = \
Firm['Code'][Firm[pred_product_code] == 1].to_list() # Firm['Code'][Firm[pred_product_code] == 1].to_list()
lst_pred_firm_size_damp = \ # lst_pred_firm_size_damp = \
[G_Firm.nodes[pred_firm]['Revenue_Log'] ** # [G_Firm.nodes[pred_firm]['Revenue_Log'] **
self.flt_netw_sply_prf_size # self.flt_netw_sply_prf_size
for pred_firm in lst_pred_firm] # for pred_firm in lst_pred_firm]
lst_prob = \ # lst_prob = \
[size_damp / sum(lst_pred_firm_size_damp) # [size_damp / sum(lst_pred_firm_size_damp)
for size_damp in lst_pred_firm_size_damp] # for size_damp in lst_pred_firm_size_damp]
# select multiple supplier (multi-sourcing) # # select multiple supplier (multi-sourcing)
n_pred_firm = self.int_netw_sply_prf_size # n_pred_firm = self.int_netw_sply_prf_n
lst_choose_firm = self.nprandom.choice(lst_pred_firm, # if n_pred_firm > len(lst_pred_firm):
n_pred_firm, # n_pred_firm = len(lst_pred_firm)
replace=False, # lst_choose_firm = self.nprandom.choice(lst_pred_firm,
p=lst_prob) # n_pred_firm,
lst_add_edge = [(pred_firm, node, # replace=False,
{'Product': pred_product_code}) # p=lst_prob)
for pred_firm in lst_choose_firm] # lst_add_edge = [(pred_firm, node,
G_Firm.add_edges_from(lst_add_edge) # {'Product': pred_product_code})
# for pred_firm in lst_choose_firm]
# G_Firm.add_edges_from(lst_add_edge)
# graph firm prod # # graph firm prod
set_node_prod_code = set(G_Firm.nodes[node]['Product_Code']) # set_node_prod_code = set(G_Firm.nodes[node]['Product_Code'])
set_pred_succ_code = set(G_bom.successors(pred_product_code)) # set_pred_succ_code = set(G_bom.successors(pred_product_code))
lst_use_pred_prod_code = list( # lst_use_pred_prod_code = list(
set_node_prod_code & set_pred_succ_code) # set_node_prod_code & set_pred_succ_code)
for pred_firm in lst_choose_firm: # for pred_firm in lst_choose_firm:
pred_node = [n for n, v in G_FirmProd.nodes(data=True) # pred_node = [n for n, v in G_FirmProd.nodes(data=True)
if v['Firm_Code'] == pred_firm and # if v['Firm_Code'] == pred_firm and
v['Product_Code'] == pred_product_code][0] # v['Product_Code'] == pred_product_code][0]
for use_pred_prod_code in lst_use_pred_prod_code: # for use_pred_prod_code in lst_use_pred_prod_code:
current_node = \ # current_node = \
[n for n, v in G_FirmProd.nodes(data=True) # [n for n, v in G_FirmProd.nodes(data=True)
if v['Firm_Code'] == node and # if v['Firm_Code'] == node and
v['Product_Code'] == use_pred_prod_code][0] # v['Product_Code'] == use_pred_prod_code][0]
G_FirmProd.add_edge(pred_node, current_node) # G_FirmProd.add_edge(pred_node, current_node)
# nx.to_pandas_adjacency(G_Firm).to_csv('adj_g_firm.csv') # nx.to_pandas_adjacency(G_Firm).to_csv('adj_g_firm.csv')
# nx.to_pandas_adjacency(G_FirmProd).to_csv('adj_g_firm_prod.csv') # nx.to_pandas_adjacency(G_FirmProd).to_csv('adj_g_firm_prod.csv')
# unconnected node # unconnected node
# for node in nx.nodes(G_Firm): for node in nx.nodes(G_Firm):
# if node. if G_Firm.degree(node) == 0:
for product_code in G_Firm.nodes[node]['Product_Code']:
# unconnect node does not have possible suppliers
lst_succ_product_code = list(
G_bom.successors(product_code))
# different from for different types of product,
# finding a common supplier (the logic above),
# for different types of product,
# finding a custormer for each product
for succ_product_code in lst_succ_product_code:
# for each product successor (finished product)
# the firm sells to,
# get a list of firm producing this finished product
lst_succ_firm = \
Firm['Code'][Firm[succ_product_code] == 1].to_list()
lst_succ_firm_size_damp = \
[G_Firm.nodes[succ_firm]['Revenue_Log'] **
self.flt_netw_cust_prf_size
for succ_firm in lst_succ_firm]
lst_prob = \
[size_damp / sum(lst_succ_firm_size_damp)
for size_damp in lst_succ_firm_size_damp]
# select multiple customer (multi-selling)
n_succ_firm = self.int_netw_cust_prf_n
if n_succ_firm > len(lst_succ_firm):
n_succ_firm = len(lst_succ_firm)
lst_choose_firm = self.nprandom.choice(lst_succ_firm,
n_succ_firm,
replace=False,
p=lst_prob)
lst_add_edge = [(node, succ_firm,
{'Product': pred_product_code})
for succ_firm in lst_choose_firm]
G_Firm.add_edges_from(lst_add_edge)
# graph firm prod
set_node_prod_code = set(
G_Firm.nodes[node]['Product_Code'])
set_pred_succ_code = set(
G_bom.successors(pred_product_code))
lst_use_pred_prod_code = list(
set_node_prod_code & set_pred_succ_code)
for pred_firm in lst_choose_firm:
pred_node = [n for n, v in G_FirmProd.nodes(data=True)
if v['Firm_Code'] == pred_firm and
v['Product_Code'] == pred_product_code][0]
for use_pred_prod_code in lst_use_pred_prod_code:
current_node = \
[n for n, v in G_FirmProd.nodes(data=True)
if v['Firm_Code'] == node and
v['Product_Code'] == use_pred_prod_code][0]
G_FirmProd.add_edge(pred_node, current_node)
self.sample.g_firm = json.dumps(nx.adjacency_data(G_Firm)) self.sample.g_firm = json.dumps(nx.adjacency_data(G_Firm))
self.firm_network = ap.Network(self, G_Firm) self.firm_network = ap.Network(self, G_Firm)

27
test.ipynb
View File

@ -239,6 +239,33 @@
" )\n", " )\n",
"print(lst_choose_firm)" "print(lst_choose_firm)"
] ]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"ename": "ValueError",
"evalue": "Cannot take a larger sample than population when replace is False",
"output_type": "error",
"traceback": [
"\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
"\u001b[1;31mValueError\u001b[0m Traceback (most recent call last)",
"Cell \u001b[1;32mIn[9], line 2\u001b[0m\n\u001b[0;32m 1\u001b[0m nprandom \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mrandom\u001b[39m.\u001b[39mdefault_rng(\u001b[39m0\u001b[39m)\n\u001b[1;32m----> 2\u001b[0m lst_choose_firm \u001b[39m=\u001b[39m nprandom\u001b[39m.\u001b[39;49mchoice([\u001b[39m1\u001b[39;49m,\u001b[39m2\u001b[39;49m],\n\u001b[0;32m 3\u001b[0m \u001b[39m3\u001b[39;49m,\n\u001b[0;32m 4\u001b[0m replace\u001b[39m=\u001b[39;49m\u001b[39mFalse\u001b[39;49;00m\n\u001b[0;32m 5\u001b[0m )\n\u001b[0;32m 6\u001b[0m lst_choose_firm\n",
"File \u001b[1;32m_generator.pyx:753\u001b[0m, in \u001b[0;36mnumpy.random._generator.Generator.choice\u001b[1;34m()\u001b[0m\n",
"\u001b[1;31mValueError\u001b[0m: Cannot take a larger sample than population when replace is False"
]
}
],
"source": [
"nprandom = np.random.default_rng(0)\n",
"lst_choose_firm = nprandom.choice([1,2],\n",
" 3,\n",
" replace=False\n",
" )\n",
"lst_choose_firm"
]
} }
], ],
"metadata": { "metadata": {