accept order conditionally

controller_db.py

@@ -16,10 +16,7 @@ class ControllerDB:
     db_name_prefix: str = None
     reset_flag: int
 
-    lst_saved_s_id_3: list
-    lst_saved_s_id_1_2: list
-
-    # n_sample_1_2: int
+    lst_saved_s_id: list
 
     def __init__(self, prefix, reset_flag=0):
         with open('conf_experiment.yaml') as yaml_file:
@@ -51,8 +48,8 @@ class ControllerDB:
             for product_code in row.index[row == 1].to_list():
                 dct = {code: [product_code]}
                 list_dct.append(dct)
-                break
-            break
+            #     break
+            # break
         # list_dct = [{'140': ['1.4.5.1']}]
         for idx_exp, dct in enumerate(list_dct):
             self.add_experiment_1(idx_exp, self.dct_parameter['n_max_trial'],

firm.py

@@ -31,7 +31,7 @@ class FirmAgent(ap.Agent):
                 # print(n1, n2, key, product_code)
                 self.firm_network.graph.remove_edge(n1, n2, key)
 
-                # remove customer up product if does not have alternative / con
+                # remove customer up product conditionally
                 customer = ap.AgentIter(self.model, n2).to_list()[0]
                 list_in_edges = list(
                     self.firm_network.graph.in_edges(n2,
@@ -126,7 +126,10 @@ class FirmAgent(ap.Agent):
             # print(product.code, [firm.name for firm in list_firm])
 
     def accept_request(self, down_firm, product):
-        if self.model.nprandom.choice([True, False], p=[0.1, 0.9]):
+        # if self.model.nprandom.choice([True, False], p=[0.1, 0.9]):
+        lst_f_s = [firm.revenue_log for firm in self.model.a_list_total_firms if product in firm.a_list_product]
+        p_accept = self.revenue_log / sum(lst_f_s)
+        if self.model.nprandom.choice([True, False], p=[p_accept, 1-p_accept]):
             self.firm_network.graph.add_edges_from([
                 (self.firm_network.positions[self],
                 self.firm_network.positions[down_firm], {

model.py

@@ -106,11 +106,20 @@ class Model(ap.Model):
                         size / sum(list_revenue_log)
                         for size in list_revenue_log
                     ]
-                    succ_firm = self.nprandom.choice(list_succ_firms,
+                    list_f_same_p = Firm['Code'][Firm[product_code] ==
+                                                   1].to_list()
+                    list_f_size_same_p = [
+                        G_Firm.nodes[f]['Revenue_Log']
+                        for f in list_f_same_p
+                    ]
+                    share = G_Firm.nodes[node]['Revenue_Log'] / sum(list_f_size_same_p)
+                    num_succ_f = round(share * len(list_succ_firms)) if round(share * len(list_succ_firms)) > 0 else 1
+                    list_choose_firm = self.nprandom.choice(list_succ_firms, num_succ_f,
                                                      p=list_prob)
+                    list_choose_firm = list(set(list_choose_firm))
                     list_added_edges = [(node, succ_firm, {
                         'Product': product_code
-                    })]
+                    }) for succ_firm in list_choose_firm]
                     G_Firm.add_edges_from(list_added_edges)
                     # print('-' * 20)
 
@@ -189,7 +198,7 @@ class Model(ap.Model):
                 firm.a_list_product_removed.append(product)
 
         # draw network
-        self.draw_network()
+        # self.draw_network()
 
     def update(self):
         self.a_list_total_firms.clean_before_time_step()

test.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 14,
    "metadata": {},
    "outputs": [
     {
@@ -20,17 +20,50 @@
       "17\n",
       "81\n"
      ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "array([2, 2])"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
     }
    ],
    "source": [
     "import numpy as np\n",
     "\n",
     "np.random.randint(0.5, 3.5)\n",
-    "p_remove = 1\n",
+    "p_remove = 0.9\n",
     "np.random.choice([True, False], p=[p_remove, 1-p_remove])\n",
     "rng = np.random.default_rng(0)\n",
     "for _ in range(10):\n",
-    "    print(rng.integers(0,100))"
+    "    print(rng.integers(0,100))\n",
+    "np.random.choice([1, 2, 3], 2, p=[0.4, 0.4, 0.2])\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "2"
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "share = 0.8\n",
+    "list_succ_firms = [1, 1]\n",
+    "round(share * len(list_succ_firms)) if round(share * len(list_succ_firms)) > 0 else 1"
    ]
   },
   {