def testInversion(self): ''' Try a simple three level SC qubit system and see if can prepare the excited state. ''' #Setup a three level qubit and a 100MHz delta Q1 = SCQubit(3, 4.987456e9, -100e6, name='Q1') systemParams = SystemParams() systemParams.add_sub_system(Q1) systemParams.add_control_ham( inphase=Hamiltonian(0.5 * (Q1.loweringOp + Q1.raisingOp)), quadrature=Hamiltonian(0.5 * (-1j * Q1.loweringOp + 1j * Q1.raisingOp))) systemParams.create_full_Ham() systemParams.measurement = Q1.levelProjector(1) #Setup the pulse parameters for the optimization pulseParams = PulseParams() pulseParams.timeSteps = 1e-9 * np.ones(30) pulseParams.rhoStart = Q1.levelProjector(0) pulseParams.rhoGoal = Q1.levelProjector(1) pulseParams.add_control_line(freq=-Q1.omega) pulseParams.H_int = Hamiltonian(Q1.omega * np.diag(np.arange(Q1.dim))) pulseParams.optimType = 'state2state' #Call the optimization optimize_pulse(pulseParams, systemParams) #Now test the optimized pulse and make sure it puts all the population in the excited state result = simulate_sequence(pulseParams, systemParams, pulseParams.rhoStart, simType='unitary')[0] assert result > 0.99
def testInversion(self): ''' Try a simple three level SC qubit system and see if can prepare the excited state. ''' #Setup a three level qubit and a 100MHz delta Q1 = SCQubit(3, 4.987456e9, -100e6, name='Q1') systemParams = SystemParams() systemParams.add_sub_system(Q1) systemParams.add_control_ham(inphase = Hamiltonian(0.5*(Q1.loweringOp + Q1.raisingOp)), quadrature = Hamiltonian(0.5*(-1j*Q1.loweringOp + 1j*Q1.raisingOp))) systemParams.create_full_Ham() systemParams.measurement = Q1.levelProjector(1) #Setup the pulse parameters for the optimization pulseParams = PulseParams() pulseParams.timeSteps = 1e-9*np.ones(30) pulseParams.rhoStart = Q1.levelProjector(0) pulseParams.rhoGoal = Q1.levelProjector(1) pulseParams.add_control_line(freq=-Q1.omega) pulseParams.H_int = Hamiltonian(Q1.omega*np.diag(np.arange(Q1.dim))) pulseParams.optimType = 'state2state' #Call the optimization optimize_pulse(pulseParams, systemParams) #Now test the optimized pulse and make sure it puts all the population in the excited state result = simulate_sequence(pulseParams, systemParams, pulseParams.rhoStart, simType='unitary')[0] assert result > 0.99
def testDRAG(self): ''' Try a unitary inversion pulse on a three level SCQuibt and see if we get something close to DRAG ''' #Setup a three level qubit and a 100MHz delta Q1 = SCQubit(3, 4.987456e9, -150e6, name='Q1') systemParams = SystemParams() systemParams.add_sub_system(Q1) systemParams.add_control_ham(inphase = Hamiltonian(0.5*(Q1.loweringOp + Q1.raisingOp)), quadrature = Hamiltonian(0.5*(-1j*Q1.loweringOp + 1j*Q1.raisingOp))) systemParams.add_control_ham(inphase = Hamiltonian(0.5*(Q1.loweringOp + Q1.raisingOp)), quadrature = Hamiltonian(0.5*(-1j*Q1.loweringOp + 1j*Q1.raisingOp))) systemParams.create_full_Ham() systemParams.measurement = Q1.levelProjector(1) #Setup the pulse parameters for the optimization numPoints = 30 pulseTime = 15e-9 pulseParams = PulseParams() pulseParams.timeSteps = (pulseTime/numPoints)*np.ones(numPoints) pulseParams.rhoStart = Q1.levelProjector(0) pulseParams.rhoGoal = Q1.levelProjector(1) pulseParams.Ugoal = Q1.pauliX pulseParams.add_control_line(freq=-Q1.omega, bandwidth=300e6, maxAmp=200e6) pulseParams.add_control_line(freq=-Q1.omega, phase=-np.pi/2, bandwidth=300e6, maxAmp=200e6) pulseParams.H_int = Hamiltonian((Q1.omega)*np.diag(np.arange(Q1.dim))) pulseParams.optimType = 'unitary' pulseParams.derivType = 'finiteDiff' #Start with a Gaussian tmpGauss = np.exp(-np.linspace(-2,2,numPoints)**2) tmpScale = 0.5/(np.sum(pulseParams.timeSteps*tmpGauss)) pulseParams.startControlAmps = np.vstack((tmpScale*tmpGauss, np.zeros(numPoints))) #Call the optimization optimize_pulse(pulseParams, systemParams) if plotResults: plt.plot(np.cumsum(pulseParams.timeSteps)*1e9,pulseParams.controlAmps.T/1e6); plt.ylabel('Pulse Amplitude (MHz)') plt.xlabel('Time (ns)') plt.legend(('X Quadrature', 'Y Quadrature')) plt.title('DRAG Pulse from Optimal Control') plt.show() #Now test the optimized pulse and make sure it does give us the desired unitary result = simulate_sequence(pulseParams, systemParams, pulseParams.rhoStart, simType='unitary') assert np.abs(np.trace(np.dot(result[1].conj().T, pulseParams.Ugoal)))**2/np.abs(np.trace(np.dot(pulseParams.Ugoal.conj().T, pulseParams.Ugoal)))**2 > 0.99
def testDRAG(self): ''' Try a unitary inversion pulse on a three level SCQuibt and see if we get something close to DRAG ''' #Setup a three level qubit and a 100MHz delta Q1 = SCQubit(3, 4.987456e9, -150e6, name='Q1') systemParams = SystemParams() systemParams.add_sub_system(Q1) systemParams.add_control_ham( inphase=Hamiltonian(0.5 * (Q1.loweringOp + Q1.raisingOp)), quadrature=Hamiltonian(0.5 * (-1j * Q1.loweringOp + 1j * Q1.raisingOp))) systemParams.add_control_ham( inphase=Hamiltonian(0.5 * (Q1.loweringOp + Q1.raisingOp)), quadrature=Hamiltonian(0.5 * (-1j * Q1.loweringOp + 1j * Q1.raisingOp))) systemParams.create_full_Ham() systemParams.measurement = Q1.levelProjector(1) #Setup the pulse parameters for the optimization numPoints = 30 pulseTime = 15e-9 pulseParams = PulseParams() pulseParams.timeSteps = (pulseTime / numPoints) * np.ones(numPoints) pulseParams.rhoStart = Q1.levelProjector(0) pulseParams.rhoGoal = Q1.levelProjector(1) pulseParams.Ugoal = Q1.pauliX pulseParams.add_control_line(freq=-Q1.omega, bandwidth=300e6, maxAmp=200e6) pulseParams.add_control_line(freq=-Q1.omega, phase=-np.pi / 2, bandwidth=300e6, maxAmp=200e6) pulseParams.H_int = Hamiltonian( (Q1.omega) * np.diag(np.arange(Q1.dim))) pulseParams.optimType = 'unitary' pulseParams.derivType = 'finiteDiff' #Start with a Gaussian tmpGauss = np.exp(-np.linspace(-2, 2, numPoints)**2) tmpScale = 0.5 / (np.sum(pulseParams.timeSteps * tmpGauss)) pulseParams.startControlAmps = np.vstack( (tmpScale * tmpGauss, np.zeros(numPoints))) #Call the optimization optimize_pulse(pulseParams, systemParams) if plotResults: plt.plot( np.cumsum(pulseParams.timeSteps) * 1e9, pulseParams.controlAmps.T / 1e6) plt.ylabel('Pulse Amplitude (MHz)') plt.xlabel('Time (ns)') plt.legend(('X Quadrature', 'Y Quadrature')) plt.title('DRAG Pulse from Optimal Control') plt.show() #Now test the optimized pulse and make sure it does give us the desired unitary result = simulate_sequence(pulseParams, systemParams, pulseParams.rhoStart, simType='unitary') assert np.abs(np.trace(np.dot( result[1].conj().T, pulseParams.Ugoal)))**2 / np.abs( np.trace(np.dot(pulseParams.Ugoal.conj().T, pulseParams.Ugoal)))**2 > 0.99
plt.plot(pulseTimes * 1e9, DRAGResults) '''State to State Optimal Control''' numSteps = 100 pulseTimes = 1e-9 * np.arange(3, 25, 3) results = np.zeros_like(pulseTimes) for ct, pulseTime in enumerate(pulseTimes): pulseParams = PulseParams() pulseParams.timeSteps = (pulseTime / numSteps) * np.ones(numSteps) pulseParams.rhoStart = qubit.levelProjector(0) pulseParams.rhoGoal = qubit.levelProjector(1) pulseParams.add_control_line(freq=0, phase=0) pulseParams.add_control_line(freq=0, phase=-pi / 2) pulseParams.type = 'state2state' #Call the optimization optimize_pulse(pulseParams, systemParams) #Now test the optimized pulse and make sure it puts all the population in the excited state results[ct] = simulate_sequence(pulseParams, systemParams, pulseParams.rhoStart, simType='lindblad')[0] plt.plot(pulseTimes * 1e9, results, '*') plt.xlabel('Pulse Time (ns)') plt.ylabel('Population in First Excited State') plt.title('Excited State Preparation with T1 = {0:.0f}ns'.format(1e9 * qubit.T1)) plt.legend(('Square', 'Gaussian', 'DRAG', 'Optimal Control')) plt.show()
'''State to State Optimal Control''' numSteps = 100 pulseTimes = 1e-9*np.arange(3,25,3) results = np.zeros_like(pulseTimes) for ct, pulseTime in enumerate(pulseTimes): pulseParams = PulseParams() pulseParams.timeSteps = (pulseTime/numSteps)*np.ones(numSteps) pulseParams.rhoStart = qubit.levelProjector(0) pulseParams.rhoGoal = qubit.levelProjector(1) pulseParams.add_control_line(freq=0, phase=0) pulseParams.add_control_line(freq=0, phase=-pi/2) pulseParams.type = 'state2state' #Call the optimization optimize_pulse(pulseParams, systemParams) #Now test the optimized pulse and make sure it puts all the population in the excited state results[ct] = simulate_sequence(pulseParams, systemParams, pulseParams.rhoStart, simType='lindblad')[0] plt.plot(pulseTimes*1e9,results,'*') plt.xlabel('Pulse Time (ns)') plt.ylabel('Population in First Excited State') plt.title('Excited State Preparation with T1 = {0:.0f}ns'.format(1e9*qubit.T1)) plt.legend(('Square','Gaussian','DRAG','Optimal Control')) plt.show()