soco 1.0.1

Algorithms for Smoothed Online Convex Optimization
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138

use crate::algorithms::online::{FractionalStep, Step};
use crate::config::{Config, FractionalConfig};
use crate::distance::NormFn;
use crate::model::{ModelOutputFailure, ModelOutputSuccess};
use crate::numerics::convex_optimization::{find_minimizer, WrappedObjective};
use crate::problem::{FractionalSmoothedConvexOptimization, Online, Problem};
use crate::result::{Failure, Result};
use crate::schedule::FractionalSchedule;
use crate::utils::{assert, sample_uniform};
use cached::{Cached, SizedCache};
use noisy_float::prelude::*;
use pyo3::prelude::*;
use serde_derive::{Deserialize, Serialize};

#[pyclass]
#[derive(Clone, Debug, Deserialize, Serialize)]
pub struct Memory {
    /// Random number $r$ representing bias.
    pub r: f64,
}
impl Default for Memory {
    fn default() -> Self {
        Memory {
            r: sample_uniform(-1., 1.),
        }
    }
}

#[pyclass]
#[derive(Clone)]
pub struct Options {
    /// Scaling factor for norm. $\theta \geq 1$. Defaults to $1$.
    #[pyo3(get, set)]
    pub theta: f64,
}
impl Default for Options {
    fn default() -> Self {
        Options { theta: 1. }
    }
}
#[pymethods]
impl Options {
    #[new]
    fn constructor(theta: f64) -> Self {
        Options { theta }
    }
}

/// Randomly Biased Greedy
pub fn rbg<C, D>(
    o: Online<FractionalSmoothedConvexOptimization<'_, C, D>>,
    mut t: i32,
    _: &FractionalSchedule,
    m: Memory,
    options: Options,
) -> Result<FractionalStep<Memory>>
where
    C: ModelOutputSuccess,
    D: ModelOutputFailure,
{
    assert(o.w == 0, Failure::UnsupportedPredictionWindow(o.w))?;
    assert(o.p.d == 1, Failure::UnsupportedProblemDimension(o.p.d))?;

    WORK.lock().unwrap().cache_clear();

    // apply lookahead
    t += 1;

    let x = next(o, t, m.r, options.theta)?;
    Ok(Step(Config::single(x), None))
}

#[derive(Clone)]
struct NextObjectiveData<'a, C, D> {
    o: Online<FractionalSmoothedConvexOptimization<'a, C, D>>,
    t: i32,
    r: f64,
    theta: f64,
}

fn next<C, D>(
    o: Online<FractionalSmoothedConvexOptimization<'_, C, D>>,
    t: i32,
    r: f64,
    theta: f64,
) -> Result<f64>
where
    C: ModelOutputSuccess,
    D: ModelOutputFailure,
{
    let bounds = o.p.bounds.clone();
    let objective = WrappedObjective::new(
        NextObjectiveData { o, t, r, theta },
        |raw_x, data| {
            let x = Config::new(raw_x.to_vec());
            w(
                &data.o.p.bounds,
                &|t, x| data.o.p.hit_cost(t, x).cost,
                &data.o.p.switching_cost,
                data.t - 1,
                data.theta,
                x.clone(),
            ) + n64(data.r) * n64(data.theta) * (data.o.p.switching_cost)(x)
        },
    );

    let (x, _) = find_minimizer(objective, bounds);
    Ok(x[0])
}

#[derive(Clone)]
struct WorkObjectiveData {
    bounds: Vec<(f64, f64)>,
    switching_cost: NormFn<f64>,
    t: i32,
    theta: f64,
    x: FractionalConfig,
}

cached_key! {
    WORK: SizedCache<String, N64> = SizedCache::with_size(1_000);
    Key = { format!("{}-{:?}", t, x) };
    fn w(bounds: &Vec<(f64, f64)>, hitting_cost: &impl Fn(i32, FractionalConfig) -> N64, switching_cost: &NormFn<f64>, t: i32, theta: f64, x: FractionalConfig) -> N64 = {
        if t == 0 {
            n64(theta) * switching_cost(x)
        } else {
            let objective = WrappedObjective::new(WorkObjectiveData { bounds: bounds.clone(), switching_cost: switching_cost.clone(), t, theta, x }, |raw_y, data| {
                let y = Config::new(raw_y.to_vec());
                w(&data.bounds, hitting_cost, &data.switching_cost, data.t - 1, data.theta, y.clone())
                    + hitting_cost(data.t, y.clone())
                    + n64(data.theta) * switching_cost(data.x.clone() - y)
            });

            let (_, opt) = find_minimizer(objective, bounds.clone());
            opt
        }
    }
}