| 研究課題/領域番号 |
23KJ0605
|
|---|
| 研究種目 |
特別研究員奨励費
|
| 配分区分 | 基金 |
|---|
| 応募区分 | 国内 |
|---|
| 審査区分 |
小区分21060:電子デバイスおよび電子機器関連
|
|---|
| 研究機関 | 東京大学 |
|---|
研究代表者 |
LIAO ZHIQIANG 東京大学, 工学系研究科, 特別研究員(PD)
|
|---|
| 研究期間 (年度) |
2023-04-25 – 2025-03-31
|
| 研究課題ステータス |
交付 (2023年度)
|
| 配分額 *注記 |
1,800千円 (直接経費: 1,800千円)
2024年度: 900千円 (直接経費: 900千円)
2023年度: 900千円 (直接経費: 900千円)
|
|---|
| キーワード | Spin wave / Asymmetric interference / Ising machine / Low power consumption / Noise assistance |
|---|
| 研究開始時の研究の概要 |
Gain-dissipative Ising machine (GIM) is an emerging device used for combinatorial optimization. In this study, the spin-wave interferometer-based GIM will be designed to realize low-power consumption, high speed, and success rate by modulating noise color and interference asymmetry.
|
| 研究実績の概要 |
(1) I fabricated a 20-nm-thick 400-μm-long Y-shape spin wave interferometer (SWI) based on yttrium iron garnet. I tested different excitation frequencies under in-plan magnetic field of 20mT. The results proved that an excitation frequency of 1.8 GHz can achieve the strongest interference asymmetry.
(2) Based on experimental data, I simulated the SWI-based Ising machine. The results proved that my device can operate under conditions where the noise intensity is 2.5 times greater than its saturation spin amplitude.
(3) To modulate the noise in the experiment, I used an arbitrary wave generator to generate the initial white noise. Then, by feeding the bias control signal and noise signal together into the comparator LT1016, statistical modulation of the noise characteristics was achieved.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
Based on the current results, I have achieved the predetermined goal. This laid the foundation and good expectations for the next-year experiments. Moreover, some related research results on noise-assisted Ising machine with asymmetric nonlinearity were also presented at the domestic conference.
|
| 今後の研究の推進方策 |
(1) Connect circuit modules (such as noise modulation circuit) with spin wave measurement systems (e.g., vector network analyzer). This can achieve the measurement and modulation of time-domain spin wave signals.
(2) Expand testing results from being limited to white noise to at least four types of colored noise conditions.
(3) Typically, achieving the highest success rate for different classes of problems (COPs) requires different spin coupling strengths and noise conditions. Therefore, in constructing hardware for spin wave GIMs, challenging COPs relevant to real-world scenarios are chosen as benchmarks. Initially, parameter ranges are roughly determined through simulations.Subsequently, optimal system parameters are determined through experiments.
|
