more improvements

src/ga_qas.py

@@ -3,24 +3,22 @@
 # "Genetic optimization of ansatz expressibility for enhanced variational quantum algorithm performance"
 
 import random
-from quantum_circuit import (
-    Gate,
-    GateType,
-    QuantumCircuit,
-    circ_from_layers,
-    sample_random_generator,
-)
+
+import matplotlib.pyplot as plt
+
 from qas_flow import Stream
+from quantum_circuit import (Gate, GateType, QuantumCircuit, circ_from_layers,
+                             sample_random_generator, single_typ)
 
-DEPTH = 20
-QUBITS = 5
-GENERATIONS = 20
-GENERATION_SIZE = 20
-PARENT_AMOUNT = 5
-MUTATION_RATE = 0.1
+DEPTH: int = 6
+QUBITS: int = 6
+GENERATIONS: int = 40
+GENERATION_SIZE: int = 60
+PARENT_AMOUNT: int = 10
+MUTATION_RATE: float = 0.1
 
 
-gate_set = [
+gate_set: list[GateType] = [
     GateType.H,
     GateType.RX,
     GateType.RY,
@@ -30,8 +28,15 @@ gate_set = [
 ]
 
 
-def main():
-    seed_rng = random.Random(1020381)
+def plot_best_circuits(best_circuits: list[QuantumCircuit]) -> None:
+    fig, ax = plt.subplots()
+
+    ax.plot([-circ.expressibility for circ in best_circuits])
+    fig.savefig("best_circuits.png")
+
+
+def main() -> None:
+    seed_rng: random.Random = random.Random(1020381)
     initial_population: list[QuantumCircuit] = (
         Stream(sample_random_generator(random.Random(101020), QUBITS, DEPTH, gate_set))
         .apply(lambda circ: print(circ))
@@ -49,6 +54,8 @@ def main():
 
     main_rng = random.Random(2837175)
 
+    best_circuits: list[QuantumCircuit] = []
+
     for generation in range(GENERATIONS):
         print(f"starting generation {generation}")
         population.sort(key=lambda qc: qc.expressibility, reverse=True)
@@ -68,22 +75,35 @@ def main():
 
                 if old_gate.single():
                     child_layers[layer_idx][gate_idx] = Gate(
-                        main_rng.choice(gate_set),
+                        main_rng.choice(
+                            [gate for gate in gate_set if single_typ(gate)]
+                        ),
                         old_gate.qubits,
                         old_gate.param_idx,
                     )
                 else:
                     child_layers[layer_idx][gate_idx] = Gate(
                         old_gate.typ,
-                        tuple(old_gate.qubits[::-1]),
+                        (old_gate.qubits[1], old_gate.qubits[0]),
                         old_gate.param_idx,
                     )
 
             child = circ_from_layers(child_layers, QUBITS)
             child.expressibility_estimate(2000, seed_rng.randint(1000, 1000000000))
             offspring.append(child)
+
+        offspring.sort(key=lambda qc: qc.expressibility, reverse=True)
+        if population[0].expressibility > offspring[0].expressibility:
+            print(f"best parent > best child")
+            best_circuits.append(population[0])
+        else:
+            print(f"best child > best parent")
+            best_circuits.append(offspring[0])
         population = offspring
 
+    plot_best_circuits(best_circuits)
+    plt.show()
+
 
 if __name__ == "__main__":
     main()

src/quantum_circuit.py

@@ -1,11 +1,12 @@
-from dataclasses import dataclass
 import math
 import random
-from typing import override
-import numpy as np
+from dataclasses import dataclass
 from enum import IntEnum
+from typing import override
 
-from qiskit import QuantumCircuit as QiskitCircuit, transpile
+import numpy as np
+from qiskit import QuantumCircuit as QiskitCircuit
+from qiskit import transpile
 from qiskit.circuit import ParameterVector, ParameterVectorElement
 from qiskit_aer import AerSimulator
 
@@ -369,11 +370,10 @@ def sample_circuit_random(
         for loc in range(qubits):
             if loc in used_qubits:
                 continue
-            gate_type = rng.choice(
-                gate_types
-                if loc + 1 < qubits
-                else [typ for typ in gate_types if single_typ(typ)]
-            )
+            if loc + 1 < qubits:
+                gate_type = rng.choice(gate_types)
+            else:
+                gate_type = rng.choice([typ for typ in gate_types if single_typ(typ)])
             match gate_type:
                 case GateType.H:
                     layer.append(Gate(gate_type, loc, params))