4 spin system protocols and spectra
收藏资源简介:
allProtocols.pkl 200 rows, corr. to 200 control sets 5 columns: cSet: nested list of control parameters. Inner list are constrained collections of parameters. 0: h0, 1: h1, 2: h2, 3: h3, 4: J01, 5: J02, 6: J03, 7: J12, 8: J13, 9: J23, 10: K012, 11: K013, 12: K023, 13: K123, 14: L u(t): (251, m) numpy arrays of protocols. Number of params is m. All protocols interpolated to 251 points for uniform data analysis V(t): (251, 16) numpy arrays of state energies over protocols Work: Linear-response excess work for tau = 1, beta = 1 Type: Labels indicating unique (up to numerical precision) control strategies. A protocol with type "P.Q" is Qth best control strategy in the collection with P independently varying parameters. spectralData.pkl 42 rows, corr. to 42 unique types 3 columns: Type: as in allProtocols.pkl Eigenvalues: (M, 15) numpy arrays of non-null eigenvalues. Most have M = 6000, others have M = 12,000 Right eigenvectors: (M, 16, 15) numpy arrays of non-null right eigenvectors NOTE: eigenvalues, eigenvectors indexed by magnitude at t = 0. Rayleigh iteration from prev. timestep was performed to preserve continuity. Only the continuity of modes that contributed non-negligibly to dissipation was confirmed, process may have failed for irrelevant modes.
allProtocols.pkl 共200行,对应200个控制集。 包含5个列: cSet:控制参数的嵌套列表,内层列表为参数的约束集合。索引0对应h0、1对应h1、2对应h2、3对应h3、4对应J01、5对应J02、6对应J03、7对应J12、8对应J13、9对应J23、10对应K012、11对应K013、12对应K023、13对应K123、14对应L。 u(t):形状为(251, m)的协议numpy数组,其中m为参数总数。所有协议均被插值至251个采样点,以支持统一的数据分析。 V(t):形状为(251, 16)的各协议下状态能量numpy数组。 Work:当τ=1、β=1时的线性响应超额功。 Type:用于标识唯一(在数值精度范围内)控制策略的标签。若某协议的类型为“P.Q”,则代表其属于包含P个独立可变参数的策略集合中排名第Q的最优控制策略。 spectralData.pkl 共42行,对应42种唯一策略类型。 包含3个列: Type:与allProtocols.pkl中的Type字段一致。 Eigenvalues:形状为(M, 15)的非零特征值numpy数组,其中多数数组的M=6000,其余数组的M=12000。 Right eigenvectors:形状为(M, 16, 15)的非零右特征向量numpy数组。 注意:特征值与特征向量按照t=0时刻的幅值进行索引。通过对上一时间步执行瑞利迭代以维持模式连续性。仅确认了对耗散具有不可忽略贡献的模式的连续性,对于无关模式,该过程可能失效。



