np.prod, head tail zuordnungs korrektur

ILP/butadien/butadiensynthesis.py

@@ -1,4 +1,4 @@
-import math
+import numpy as np
 import gurobipy as gp
 from gurobipy import GRB, Model, quicksum
 
@@ -91,15 +91,15 @@ VERTICES = ['Butadien', 'p_{0,0}', 'p_{0,1}', 'p_{0,2}', 'p_{0,3}', 'p_{0,4}', '
 #Vergleich mit dem NMR von Ethylen und Hexatrien
 NMR1 = [0.32, 0.5, 0.87, 0.0, 0.11, 0.58, 0.06, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.09, 0.0, 0.0, 0.17, 0.0, 0.09]
 #Normalisiert
-NMR1 = [0.11, 0.18, 0.31, 0.0, 0.04, 0.21, 0.02, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03, 0.0, 0.0, 0.06, 0.0, 0.03]
+#NMR1 = [0.11, 0.18, 0.31, 0.0, 0.04, 0.21, 0.02, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03, 0.0, 0.0, 0.06, 0.0, 0.03]
 #Vergleich mit dem NMR von Ethylen und Octrien
 NMR2 = [0.33, 0.43, 0.6, 0.0, 0.13, 0.9, 0.07, 0.01, 0.0, 0.0, 0.01, 0.0, 0.0, 0.01, 0.1, 0.0, 0.0, 0.22, 0.0, 0.13]
 #Normalisiert
-NMR2 = [0.11, 0.15, 0.21, 0.0, 0.04, 0.31, 0.02, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03, 0.0, 0.0, 0.07, 0.0, 0.04]
+#NMR2 = [0.11, 0.15, 0.21, 0.0, 0.04, 0.31, 0.02, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03, 0.0, 0.0, 0.07, 0.0, 0.04]
 #Vergleich mit dem NMR von Ethylen und Benzol
 NMR3 = [0.0, 0.32, 0.0, 0.0, 0.0, 0.0, 0.05, 0.0, 0.0, 0.02, 0.0, 0.29, 0.95, 0.17, 0.0, 0.0, 0.0, 0.0, 0.0, 0.05]
 #Normalisiert
-NMR3 = [0.0, 0.17, 0.0, 0.0, 0.0, 0.0, 0.03, 0.0, 0.0, 0.01, 0.0, 0.16, 0.51, 0.09, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03]
+#NMR3 = [0.0, 0.17, 0.0, 0.0, 0.0, 0.0, 0.03, 0.0, 0.0, 0.01, 0.0, 0.16, 0.51, 0.09, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03]
 VERTICESSMILES = ['C=CC=C', 'C=C', 'C=CC=CC=C', 'C1CCC(C=C)CC=1', 'C(CCC(C=C)CC=C)=C', 'C=CC=CC=CC=C', 'C1CCC(C=CC=C)CC=1', 'C=CC1C=CCCC1', 'C1CCCCC=1', 'C=CC1CC=CCC1C=C', 'C1CCC(C=C)C(C=C)C=1', 'C(C1CC(C=C)C=CC1)=C', 'C1C=CC=CC=1', 'C(CCC1C=CC=CC1)=C', 'C1C(C=CC=C)CCCC=1', 'C=CC(C=C)CCCC=C', 'C=CCCCCC=C', 'C=CC1C=CC(C=C)CC1', 'C1CC2CCCCC2CC=1', 'C1CC2C=CC=CC2CC=1']
 NMRMAX = [max(n1,n2,n3) for n1,n2,n3 in zip(NMR1,NMR2,NMR3)]
 
@@ -127,22 +127,30 @@ EDGEMAX = {}
 
 modes = ["Product", "Average"]
 mode = modes[0]
+
+#print(round(np.prod([1, 0.9])))
 #Ordnet die Kantenwahrscheinlichkeit basierend auf dem Mittel oder Produkt der Wahrscheinlichkeiten der Produkte
-for edge, (heads, tails) in HYPERGRAPH.items():
+for edge, (tails, heads) in HYPERGRAPH.items():
-    if heads != [] and tails != [] and mode == "Product":
+    if heads == [] or tails == []:
-        #Produkt klappt nur, wenn normalisierte Likelihoods
-        EDGE1[edge] = math.prod(VERTICE1[head] for head in heads)
-        EDGE2[edge] = math.prod(VERTICE2[head] for head in heads)
-        EDGE2[edge] = math.prod(VERTICE2[head] for head in heads)
-    if heads != [] and tails != [] and mode == "Average":
-        EDGE1[edge] = quicksum(VERTICE1[head] for head in heads)/len(heads)
-        EDGE2[edge] = quicksum(VERTICE2[head] for head in heads)/len(heads)
-        EDGE3[edge] = quicksum(VERTICE3[head] for head in heads)/len(heads)
-    else:
         EDGE1[edge] = 0.0
         EDGE2[edge] = 0.0
         EDGE3[edge] = 0.0
-    EDGEMAX[edge] = max([EDGE1[edge], EDGE2[edge], EDGE2[edge]])
+    elif mode == "Product":
+        #Produkt klappt nur, wenn normalisierte Likelihoods
+        #print([VERTICE1[head] for head in heads])
+        EDGE1[edge] = round(np.prod([VERTICE1[head] for head in heads]),2)
+        EDGE2[edge] = round(np.prod([VERTICE2[head] for head in heads]),2)
+        EDGE3[edge] = round(np.prod([VERTICE3[head] for head in heads]),2)
+    elif mode == "Average":
+        print(tails)
+        EDGE1[edge] = round(sum(VERTICE1[head] for head in heads)/len(heads),2)
+        EDGE2[edge] = round(sum(VERTICE2[head] for head in heads)/len(heads),2)
+        EDGE3[edge] = round(sum(VERTICE3[head] for head in heads)/len(heads),2)
+
+    EDGEMAX[edge] = max([EDGE1[edge], EDGE2[edge], EDGE3[edge]])
+    print(EDGEMAX[edge], HYPERGRAPH[edge])
+#print(EDGEMAX[4])
+#print(EDGEMAX[42])
 
 
 FIXED_FLOWS = {
@@ -180,8 +188,8 @@ def build_model(name, hyperedges, edgelikelihoodsmax, edgelikelihoods1, edgelike
         count[v] = 0
     count = model.addVars(vertices, vtype=GRB.CONTINUOUS, lb = 0.0, name = "count")
     for mol in vertices:
-        for edge, (heads, tails) in hyperedges.items():
+        for edge, (tails, heads) in hyperedges.items():
-            count[mol] += x[edge] * (heads.count(mol) - tails.count(mol))
+            count[mol] += x[edge] * (tails.count(mol) - heads.count(mol))
         model.addConstr(count[mol] == 0)
     '''
 
@@ -209,7 +217,7 @@ def build_model(name, hyperedges, edgelikelihoodsmax, edgelikelihoods1, edgelike
     #Multiply edgelikelihood with the edge use boolean
     #Adapt to have 
     model.setObjectiveN(
-        quicksum(enmax[e_id] * b[e_id] for e_id, (_, _) in hyperedges.items()),
+        quicksum(enmax[e_id] * b[e_id] for e_id in hyperedges),
         index = 0,
         priority = 2,
         name = "maximize_nmr_similarity",
@@ -223,20 +231,20 @@ def build_model(name, hyperedges, edgelikelihoodsmax, edgelikelihoods1, edgelike
     )
     
     #Excluding creation and destruction only three reactions for three nmr
-    model.addConstr(quicksum(b[e_id] for e_id, (heads, tails) in hyperedges.items() if tails != [] and heads != []) == 3) 
+    model.addConstr(quicksum(b[e_id] for e_id, (tails, heads) in hyperedges.items() if tails != [] and heads != []) == 3) 
     #Restrict number of used edges to prevent using all available
     #model.addConstr(quicksum(x[e_id] for e_id in hyperedges) <= 16)
 
     #2 Butadien create first different molecule and it has to be created first:
     model.addConstr(b[4] + b[7] == 1)
     model.addConstr(b[213] == 1)
-    model.addConstr(b[41] == 1)
+    #model.addConstr(b[41] == 1)
 
 
 
     #No cyclic reaction pairs:
-    for e_id1, (heads1, tails1) in hyperedges.items():
+    for e_id1, (tails1, heads1) in hyperedges.items():
-        for e_id2, (heads2, tails2) in hyperedges.items():
+        for e_id2, (tails2, heads2) in hyperedges.items():
             if (heads1, tails1) == (tails2, heads2):
                 #print(e_id1, e_id2)
                 model.addConstr((b[e_id1] + b[e_id2]) <= 1)
@@ -247,12 +255,16 @@ def build_model(name, hyperedges, edgelikelihoodsmax, edgelikelihoods1, edgelike
 def positive_entries(variable_dict, threshold = 0.5):
     return {e_id: var.X for e_id, var in variable_dict.items() if var.X > threshold}
 
-def print_solution(title, flow_solution, binary_solution, hyperedges):
+def edgelikelihoods(variable_dict, threshold = 0.5):
+    return {e_id: EDGEMAX[e_id] for e_id, var in variable_dict.items() if var.X > threshold}
+
+def print_solution(title, flow_solution, binary_solution, edge_likelihoods, hyperedges):
     print(f"\n{title}:")
     for e_id in sorted(flow_solution):
         flow = flow_solution[e_id]
         tails, heads = hyperedges[e_id]
-        print(f"Hyperedge {e_id}: Flow = {flow}, Tails = {tails}, Heads = {heads}")
+        likelihood = edge_likelihoods[e_id]
+        print(f"Hyperedge {e_id}: Flow = {flow}, Tails = {tails}, Heads = {heads}, With likelihood = {likelihood}")
     print("\nBinary Variables:")
     for e_id in sorted(binary_solution):
         print(f"Binary Variable b_{e_id} = {binary_solution[e_id]}")
@@ -268,7 +280,8 @@ def main():
         return
     optimal_solution = positive_entries(x)
     optimal_binary_solution = positive_entries(b)
-    print_solution("Optimal Solution", optimal_solution, optimal_binary_solution, HYPERGRAPH)
+    optimal_edgelikelihoods = edgelikelihoods(b)
+    print_solution("Optimal Solution", optimal_solution, optimal_binary_solution, optimal_edgelikelihoods, HYPERGRAPH)
 
     excluded_support = list(optimal_binary_solution.keys())
     second_model, x2, b2 = build_model("SecondBestHypergraphFlow", HYPERGRAPH, EDGEMAX, EDGE1, EDGE2, EDGE3, excluded_support=excluded_support,)
@@ -277,9 +290,9 @@ def main():
     if second_model.status == GRB.Status.OPTIMAL:
         second_solution = positive_entries(x2)
         second_binary_solution = positive_entries(b2)
-        print_solution("Second Best Solution", second_solution, second_binary_solution, HYPERGRAPH)
+        second_edgelikelihoods = edgelikelihoods(b2)
+        print_solution("Second Best Solution", second_solution, second_binary_solution, second_edgelikelihoods, HYPERGRAPH)
     else:
         print("No optimal solution found for the second best model.")
-
 if __name__ == "__main__":
     main()