quantrs2_device/vqa_support/
statistical.rs

1//! Statistical analysis and validation tools for VQA
2//!
3//! This module provides statistical methods for analyzing VQA convergence,
4//! parameter distributions, and optimization landscapes.
5
6use scirs2_core::ndarray::{Array1, Array2};
7use std::collections::HashMap;
8
9/// Statistical analysis results for VQA optimization
10#[derive(Debug, Clone)]
11pub struct VQAStatistics {
12    /// Convergence metrics
13    pub convergence_rate: f64,
14    /// Parameter variance across iterations
15    pub parameter_variance: Vec<f64>,
16    /// Objective function statistics
17    pub objective_stats: ObjectiveStatistics,
18}
19
20/// Objective function statistical metrics
21#[derive(Debug, Clone)]
22pub struct ObjectiveStatistics {
23    /// Mean objective value
24    pub mean: f64,
25    /// Standard deviation
26    pub std: f64,
27    /// Best achieved value
28    pub best: f64,
29    /// Worst value encountered
30    pub worst: f64,
31}
32
33impl Default for VQAStatistics {
34    fn default() -> Self {
35        Self {
36            convergence_rate: 0.0,
37            parameter_variance: Vec::new(),
38            objective_stats: ObjectiveStatistics::default(),
39        }
40    }
41}
42
43impl Default for ObjectiveStatistics {
44    fn default() -> Self {
45        Self {
46            mean: 0.0,
47            std: 0.0,
48            best: 0.0,
49            worst: 0.0,
50        }
51    }
52}
53
54/// Analyze VQA convergence statistics
55pub fn analyze_convergence(objective_history: &[f64]) -> VQAStatistics {
56    if objective_history.is_empty() {
57        return VQAStatistics::default();
58    }
59
60    let mean = objective_history.iter().sum::<f64>() / objective_history.len() as f64;
61    let variance: f64 = objective_history
62        .iter()
63        .map(|x| (x - mean).powi(2))
64        .sum::<f64>()
65        / objective_history.len() as f64;
66    let std = variance.sqrt();
67
68    let best = objective_history
69        .iter()
70        .fold(f64::INFINITY, |a, &b| a.min(b));
71    let worst = objective_history
72        .iter()
73        .fold(f64::NEG_INFINITY, |a, &b| a.max(b));
74
75    VQAStatistics {
76        convergence_rate: if objective_history.len() > 1 {
77            (objective_history[0] - objective_history[objective_history.len() - 1]).abs()
78                / objective_history.len() as f64
79        } else {
80            0.0
81        },
82        parameter_variance: vec![variance],
83        objective_stats: ObjectiveStatistics {
84            mean,
85            std,
86            best,
87            worst,
88        },
89    }
90}
quantrs2_device/vqa_support/statistical.rs

quantrs2_device/vqa_support/
statistical.rs
