1: %#!platex journal.tex
2: %%
3: %% analysis_first.tex %%
4: %%
5: %% Last modified: 05/03/25 20:33:36 on dell148
6: %%
7:
8: \subsection{Decoherence on the first register}
9: \label{sec:analysis_first}
10:
11: \begin{figure}[t]
12: \begin{center}
13: \includegraphics{figures/position_error1st.eps} \\
14: %\scalebox{0.525}{\includegraphics{figures/position_error1st.eps}}
15: \end{center}
16: % \vspace{-4zw}
17: \caption[Position of $\sigma_i$ error on the first register]{
18: $\sigma_i$ error occurs
19: on the $j$-th qubit of the first register
20: after application of controlled-$G^k$ operations with the $j$-th control qubit.
21: }
22: \label{fig:position_error1st}
23: \end{figure}
24:
25: %In this subsection,
26: We analyze the decoherence on the first register
27: by considering the case
28: that the error based on the depolarizing channel
29: occurs once on the first register.
30: Through the depolarizing channel,
31: $\sigma_x, \sigma_y, \sigma_z$ errors
32: and identity operator $\sigma_0$ occur
33: with the same rate
34: from Equation~(\ref{eq:depolarizing}).
35: We first deal with the ascending-order circuit
36: in Figure~\ref{fig:two_implementations}~(a).
37:
38: We investigate influences of the decoherence
39: by calculating probability distributions on the quantum counting.
40: %In order to calculate it,
41: For calculations,
42: we need to represent the position of $\sigma_i$ error.
43: %By Assumption~\ref{asmp:position},
44: As stated above,
45: we already restrict the position of the decoherence
46: on the state~(\ref{enu:counting_3}).
47: Let $j$ and $k$ be integers such that
48: the error occurs
49: \begin{itemize}
50: \item on the $j$-th qubit of the first register,
51: \item after application of controlled-$G^k$ operations
52: with the $j$-th control qubit,
53: \end{itemize}
54: where $0 \le j \le p-1, \ 0 \le k \le 2^j$.
55: These parameters are sufficient for determining
56: where and when $\sigma_i$ error occurs.
57: The case that
58: the error occurs on the $j$-th qubit
59: before(after) controlled-$G^{2^j}$
60: can be represented by $k=0 (k=2^j)$ respectively
61: since our decoherence model is local.
62: Figure~\ref{fig:position_error1st} shows
63: the position of $\sigma_i$ error on our analyses.
64: The total number of applications
65: of controlled-$G$ operations
66: before the error
67: is $2^j+k-1$.
68:
69: We first focus on the position of peaks
70: in a probability distribution on the quantum counting
71: under the decoherence.
72: If no error occurs,
73: the probability distribution has
74: only two correct peaks
75: near $\tilde{m} \simeq f, 2^p-f$,
76: as shown in Equation~(\ref{eq:counting_noerror}).
77: %\begin{proof}
78: Let $Prob^{(i,j,k)} (m')$ be
79: the probability to obtain $m'$ as a measurement result
80: in the case of the above position of $\sigma_i$ error.
81: % in the case that $\sigma_i$ error occurs on the $j$-th qubit
82: % of the first register
83: % after application of controlled-$G^k$ operations
84: % with the $j$-th control qubit.
85: By calculation $Prob^{(i,j,k)} (m')$
86: in~\ref{sec:appendix_first},
87: we have
88: \begin{eqnarray}
89: \sum_{i=0, x, y, z} Prob^{(i, j, k)}(m') & = & \hspace{31mm} \nonumber \\
90: & & \hspace{-4cm}
91: \Biggl[
92: \frac{\sin \left\{ \pi ( m' + f ) \right\} }
93: {2^{p-j-1} \sin \left\{ \pi/ 2^{p-j-1}( m' + f ) \right\} } \nonumber
94: %& & \hspace{-2.5cm}
95: \mbox{} \times
96: \frac{2^{p-j} \sin \left\{ \pi / 2^{p-j} ( m' + f ) \right\} }
97: {2^{p} \sin \left\{ \pi / 2^{p} ( m' + f ) \right\} }
98: \Biggr]^2 \nonumber \\
99: & & \hspace{-4cm}
100: \mbox{}+
101: \Biggl[
102: \frac{\sin \pi ( m' - f )}
103: {2^{p-j-1} \sin \left\{ \pi/ 2^{p-j-1}( m' - f ) \right\} }
104: %& & \hspace{-2.5cm}
105: \mbox{} \times
106: \frac{2^{p-j} \sin \left\{ \pi / 2^{p-j} ( m' - f ) \right\} }
107: {2^{p} \sin \left\{ \pi / 2^{p} ( m' - f ) \right\} }
108: \Biggr]^2. \label{eq:prob_ijk}
109: \end{eqnarray}
110: %
111: This equation has two strong peaks at $m'\simeq f, -f\equiv 2^p-f$
112: that are the same as the correct peaks
113: and weak peaks at a distance of $\pm 2^{p-j-1}$ from the strong peaks,
114: which is caused by errors.
115: \begin{proposition}
116: \label{thm:positions_wrong_first}
117: The probability distribution
118: related to the first order term of
119: the error rate
120: on the quantum counting
121: mainly has wrong peaks
122: at a distance of the power of two from $\tilde{m} \simeq f, 2^p - f$
123: if the depolarizing channel is applied
124: on the first register.
125: \end{proposition}
126:
127: \begin{figure}[t]
128: % \vspace{-2mm}
129: \begin{center}
130: \begin{minipage}{0.45\linewidth}
131: \begin{center}
132: \hspace{-10mm}
133: %\scalebox{0.87}{\includegraphics{figures/err1exp.eps}} \\
134: \scalebox{0.87}{\includegraphics{figures/error_first_m.eps}} \\
135: \vspace{-2mm}\hspace{4mm}\scalebox{1.00}{$m'$}\\
136: \mbox{\hspace{3mm} (a)}
137: %\vspace{2mm}
138: \end{center}
139: \end{minipage}
140: %\hspace{2mm}
141: \begin{minipage}{0.45\linewidth}
142: \begin{center}
143: %\vspace{-1.5zw}
144: %\scalebox{0.87}{\includegraphics{figures/err1sim.eps}} \\
145: \scalebox{1.0}{\includegraphics[angle=270,width=\textwidth]{figures/6_5_100000_upper_l_afterG_1d3_m.ps}}
146: %\includegraphics[width=\textwidth]{figures/6_5_100000_upper_l_afterG_1d3_m.eps} \\
147: %\includegraphics{figures/6_5_100000_upper_l_afterG_1d3_m.eps} \\
148:
149: \hspace{1mm}
150: \vspace{0.5mm}
151: measurement result $\tilde{m}$ \\
152: \hspace{1mm}(b)
153: %\vspace{2mm}
154: \end{center}
155: \end{minipage}
156: %\vspace{-2mm}
157: \caption[Graphs of measurement result in the first register case]{
158: The probability distributions of measurement results
159: with the same parameters in Figure~\ref{fig:no_error_m}.
160: (a):~$\frac{1}{4p} \sum_{j=0}^{p-1} \sum_{i=0,x,y,z}
161: Prob^{(i,j,k)}(m')$.
162: (b): the average of numerical calculations $10^5$ trials
163: where the error rate $d=4 \times 10^{-3}$.
164: }
165: \label{fig:err1}
166: \end{center}
167: \end{figure}
168: %\vspace{-3mm}
169:
170: %\begin{figure}
171: % % \vspace{-2mm}
172: % \begin{center}
173: % \begin{minipage}{0.45\linewidth}
174: % \begin{center}
175: % \hspace{-10mm}
176: % \scalebox{0.87}{\includegraphics{figures/error_first_t.eps}}
177: %
178: % \vspace{-1.0zw}
179: % $t'$ \\
180: % (a)
181: % \vspace{1.0zw}
182: %% \scalebox{0.87}{\includegraphics{figures/err1exp.eps}} \\
183: %% \vspace{-2mm}\hspace{4mm}\scalebox{1.00}{$m'$}\\
184: %% \mbox{\hspace{3mm} (a)}
185: %% \vspace{2mm}
186: % \end{center}
187: % \end{minipage}
188: % \hspace{2mm}
189: % \begin{minipage}{0.45\linewidth}
190: % \begin{center}
191: % \vspace{-2zw}
192: % \scalebox{1.0}{\includegraphics[angle=270,width=\textwidth]{figures/6_5_100000_upper_l_afterG_1d3_t.ps}}
193: %
194: % output $\tilde{t}$ \\
195: % (b)
196: %% \scalebox{0.87}{\includegraphics{figures/err1sim.eps}} \\
197: %% \vspace{-2mm}\hspace{4mm}\scalebox{1.00}{$m'$}\\
198: %% \mbox{\hspace{3mm} (b)}
199: %% \vspace{2mm}
200: % \end{center}
201: % \end{minipage}
202: % \vspace{-2mm}
203: % \caption[Graphs of output in the first register case]{
204: % Graphs of probability distributions of the outputs.
205: % (a) and (b) corresponds to Figure~\ref{fig:err1}~(a) and (b),
206: % respectively.
207: % }
208: % \label{fig:err1_outputs}
209: % \end{center}
210: % \vspace{-3mm}
211: %\end{figure}
212:
213: We show the graph of
214: $\frac{1}{4p} \sum_{i=0,x,y,z} \sum_{j=0}^{p-1} Prob^{(i,j,k)}(m')$
215: in Figure~\ref{fig:err1}~(a),
216: which means the average of probability distributions
217: in all error cases.
218: We also show the graph of numerical calculations
219: in Figure~\ref{fig:err1}~(b).
220: We did the experiments on Quantum Computation Simulation System
221: (QCSS)~\cite{NMI02simulator}
222: and took the average of $10^5$ trials.
223: On the numerical calculations,
224: we set the error rate $d= 4\times 10^{-3}$
225: so that the decoherence errors
226: based on the depolarizing channel
227: occur on the first register about twice on each trial.
228: %Even if the errors occur more than once,
229: %main wrong peaks appear at a distance
230: %of the power of two from two strong correct peaks.
231:
232: Proposition~\ref{thm:positions_wrong_first}
233: follows the number of wrong peaks caused by the decoherence.
234:
235: \begin{claim}
236: The number of main wrong peaks is $O(p)$
237: in probability distribution
238: related to the first order term of
239: the error rate
240: if the depolarizing channel is applied
241: on the first register.
242: \end{claim}
243:
244: %\begin{proof}
245: % The number of wrong peaks
246: % at a distance of $\pm 2^{p-j-1}$
247: % from $m' \simeq f$
248: %
249: % Proposition~\ref{thm:positions_wrong_first}
250: % states that
251: % wrong peaks mainly appear at a distance of $\pm 2^{p-j-1}$
252: % from correct peaks.
253: % It means that the number of wrong peaks
254: % from $m' \simeq f$ is at most $p$,
255: % that is the total number of wrong peaks is at most $2p=O(n)$.
256: %\end{proof}
257:
258: The probability distribution
259: $\sum_{i=0,x,y,z} Prob^{(i,j,k)}(m')$ does not
260: depend on $k$ that determines {\it depth} of $\sigma_i$ error,
261: i.e.{}\ the time when the error occurs.
262:
263: %% もうひとつの重要な定理
264: \begin{proposition}
265: \label{thm:independence_first}
266: The probability distribution
267: related to the first order term of
268: the error rate
269: on the quantum counting
270: is independent of depth of error
271: if the depolarizing channel is applied
272: on the first register.
273: \end{proposition}
274:
275: Proposition~\ref{thm:independence_first}
276: means that there is no difference
277: of influences of the decoherence
278: between on the ascending-order circuit
279: and on the descending-order one for the quantum counting,
280: %Proposition~\ref{thm:independence_first} means that
281: %the depolarizing channel is considered
282: %to be commutative to controlled-Grover operators
283: %in the case that the error occurs once on the first register.
284: since controlled-Grover operators are
285: commutative each other.%, we obtain the following proposition.
286:
287: %%% エラーの影響は，quantum counting の実装方法によらない．
288: \begin{proposition}
289: \label{thm:independence_implementation}
290: Probability distributions
291: related to the first order term of
292: the error rate
293: on the quantum counting
294: % Influences of the decoherence
295: % related to the first order term
296: are independent of the ordering of
297: application of controlled-$G$ operations
298: if the depolarizing channel is applied
299: on the first register.
300: \end{proposition}
301:
302: