ripopt 0.7.0

A memory-safe interior point optimizer in Rust
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
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164

use ripopt::{NlpProblem, SolverOptions};

/// HS TP020: Rosenbrock-like with 3 nonlinear inequality constraints
///
/// min  f(x) = 100*(x2 - x1^2)^2 + (1 - x1)^2
///            = -2*x1 + 100*x1^4 + 100*x2^2 - 200*x2*x1^2 + 1 + x1^2
///
/// s.t. g1: x1 + x2^2 >= 0
///      g2: x1^2 + x2 >= 0
///      g3: x1^2 + x2^2 >= 1
///
///      -0.5 <= x1 <= 0.5
///      x2 free
///
/// Starting point: x0 = (0.1, 1.0)
pub struct HsTp020;

impl NlpProblem for HsTp020 {
    fn num_variables(&self) -> usize { 2 }
    fn num_constraints(&self) -> usize { 3 }

    fn bounds(&self, x_l: &mut [f64], x_u: &mut [f64]) {
        x_l[0] = -0.5; x_u[0] = 0.5;
        x_l[1] = f64::NEG_INFINITY; x_u[1] = f64::INFINITY;
    }

    fn constraint_bounds(&self, g_l: &mut [f64], g_u: &mut [f64]) {
        g_l[0] = 0.0; g_u[0] = f64::INFINITY;
        g_l[1] = 0.0; g_u[1] = f64::INFINITY;
        g_l[2] = 0.0; g_u[2] = f64::INFINITY;
    }

    fn initial_point(&self, x0: &mut [f64]) {
        x0[0] = 0.1; x0[1] = 1.0;
    }

    fn objective(&self, x: &[f64], _new_x: bool, obj: &mut f64) -> bool {
        *obj = -2.0*x[0] + 100.0*x[0].powi(4) + 100.0*x[1].powi(2) - 200.0*x[1]*x[0].powi(2) + 1.0 + x[0].powi(2);
        true
    }

    fn gradient(&self, x: &[f64], _new_x: bool, grad: &mut [f64]) -> bool {
        grad[0] = 2.0*x[0] + 400.0*x[0].powi(3) - 400.0*x[0]*x[1] - 2.0;
        grad[1] = -200.0*x[0].powi(2) + 200.0*x[1];
        true
    }

    fn constraints(&self, x: &[f64], _new_x: bool, g: &mut [f64]) -> bool {
        g[0] = x[0] + x[1].powi(2);
        g[1] = x[0].powi(2) + x[1];
        g[2] = -1.0 + x[0].powi(2) + x[1].powi(2);
        true
    }

    fn jacobian_structure(&self) -> (Vec<usize>, Vec<usize>) {
        (vec![0, 0, 1, 1, 2, 2], vec![0, 1, 0, 1, 0, 1])
    }

    fn jacobian_values(&self, x: &[f64], _new_x: bool, vals: &mut [f64]) -> bool {
        vals[0] = 1.0;       vals[1] = 2.0*x[1];
        vals[2] = 2.0*x[0];  vals[3] = 1.0;
        vals[4] = 2.0*x[0];  vals[5] = 2.0*x[1];
        true
    }

    fn hessian_structure(&self) -> (Vec<usize>, Vec<usize>) {
        (vec![0, 1, 1], vec![0, 0, 1])
    }

    fn hessian_values(&self, x: &[f64], _new_x: bool, obj_factor: f64, lambda: &[f64], vals: &mut [f64]) -> bool {
        vals[0] = obj_factor * (1200.0*x[0].powi(2) - 400.0*x[1] + 2.0) + lambda[1]*2.0 + lambda[2]*2.0;
        vals[1] = obj_factor * (-400.0*x[0]);
        vals[2] = obj_factor * 200.0 + lambda[0]*2.0 + lambda[2]*2.0;
        true
    }
}

fn main() {
    env_logger::init();

    let problem = HsTp020;

    println!("=== TP020: Rosenbrock with 3 nonlinear inequality constraints ===");
    println!("min  100*(x2 - x1^2)^2 + (1-x1)^2");
    println!("s.t. g1: x1 + x2^2 >= 0");
    println!("     g2: x1^2 + x2 >= 0");
    println!("     g3: x1^2 + x2^2 >= 1");
    println!("     -0.5 <= x1 <= 0.5");
    println!("x0 = (0.1, 1.0)");
    println!();

    // Evaluate at starting point
    let x0 = [0.1, 1.0];
    let mut f0 = 0.0; problem.objective(&x0, true, &mut f0);
    let mut g0 = [0.0; 3];
    problem.constraints(&x0, true, &mut g0);
    let mut grad0 = [0.0; 2];
    problem.gradient(&x0, true, &mut grad0);
    println!("At x0: f = {:.6}, g = [{:.6}, {:.6}, {:.6}]", f0, g0[0], g0[1], g0[2]);
    println!("Gradient at x0: [{:.6}, {:.6}]", grad0[0], grad0[1]);
    println!("Constraint satisfaction: g1={} g2={} g3={}",
             if g0[0] >= 0.0 { "OK" } else { "VIOLATED" },
             if g0[1] >= 0.0 { "OK" } else { "VIOLATED" },
             if g0[2] >= 0.0 { "OK" } else { "VIOLATED" });

    // Evaluate the Hessian eigenvalues at x0 to see if it's indefinite
    let mut hvals = [0.0; 3];
    problem.hessian_values(&x0, true, 1.0, &[0.0, 0.0, 0.0], &mut hvals);
    // H = [[h00, h10], [h10, h11]] = [[hvals[0], hvals[1]], [hvals[1], hvals[2]]]
    let h00 = hvals[0];
    let h10 = hvals[1];
    let h11 = hvals[2];
    let trace = h00 + h11;
    let det = h00 * h11 - h10 * h10;
    let disc = (trace * trace - 4.0 * det).max(0.0).sqrt();
    let eig1 = (trace + disc) / 2.0;
    let eig2 = (trace - disc) / 2.0;
    println!("\nHessian at x0 (obj only): H = [[{:.2}, {:.2}], [{:.2}, {:.2}]]", h00, h10, h10, h11);
    println!("Hessian eigenvalues: {:.4}, {:.4}", eig1, eig2);
    println!("Hessian is {}", if eig2 >= 0.0 { "PSD" } else { "INDEFINITE" });

    // Check what happens at the solution x* = (0.5, ~0.25)
    // The known optimal for TP020 is approximately x* = (0.5, 0.25), f* = 0.25 (bound active)
    let x_star = [0.5, 0.25];
    let mut f_star = 0.0; problem.objective(&x_star, true, &mut f_star);
    let mut g_star = [0.0; 3];
    problem.constraints(&x_star, true, &mut g_star);
    let mut grad_star = [0.0; 2];
    problem.gradient(&x_star, true, &mut grad_star);
    println!("\nAt x*=(0.5, 0.25): f = {:.6}, g = [{:.6}, {:.6}, {:.6}]", f_star, g_star[0], g_star[1], g_star[2]);
    println!("Gradient at x*: [{:.6}, {:.6}]", grad_star[0], grad_star[1]);
    println!("Constraint satisfaction at x*: g1={} g2={} g3={}",
             if g_star[0] >= 0.0 { "OK" } else { "VIOLATED" },
             if g_star[1] >= 0.0 { "OK" } else { "VIOLATED" },
             if g_star[2] >= -1e-10 { "OK" } else { "VIOLATED" });
    let mut hvals_star = [0.0; 3];
    problem.hessian_values(&x_star, true, 1.0, &[0.0, 0.0, 0.0], &mut hvals_star);
    let h00s = hvals_star[0];
    let h10s = hvals_star[1];
    let h11s = hvals_star[2];
    let traces = h00s + h11s;
    let dets = h00s * h11s - h10s * h10s;
    let discs = (traces * traces - 4.0 * dets).max(0.0).sqrt();
    println!("Hessian eigenvalues at x*: {:.4}, {:.4}", (traces + discs)/2.0, (traces - discs)/2.0);
    println!();

    // Now run the solver
    let mut options = SolverOptions::default();
    options.print_level = 5;
    options.max_iter = 100;

    println!("=== Running solver ===\n");
    let result = ripopt::solve(&problem, &options);

    println!("\n=== Final Results ===");
    println!("Status: {:?}", result.status);
    println!("Objective: {:.10}", result.objective);
    println!("Solution: {:?}", result.x);
    println!("Iterations: {}", result.iterations);
    println!("Constraint values: {:?}", result.constraint_values);
    println!("Multipliers (y): {:?}", result.constraint_multipliers);
    println!("Bound mult lower: {:?}", result.bound_multipliers_lower);
    println!("Bound mult upper: {:?}", result.bound_multipliers_upper);
}