hdim-opt 1.0.8

Optimization toolkit for high-dimensional, non-differentiable problems.

hdim-opt: High-Dimensional Optimization Toolkit

A modern optimization suite for complex, high-dimensional problems. This package provides algorithms to accelerate convergence, including the QUASAR evolutionary algorithm and HDS non-uniform QMC sampler.

All core functions, listed below, are single-line executable and require only three essential parameters: [obj_function, bounds, n_samples].

• quasar: QUASAR optimization.
• hds: Generate a non-uniform HDS sample sequence.
• sensitivity: Perform a global Sobol sensitivity analysis (via SALib).
• sobol: Generate a uniform sample sequence (via SciPy).

---

Installation

Installed via hdim_opt directly from PyPI:

pip install hdim_opt

QUASAR Optimizer (Quasi-Adaptive Search with Asymptotic Reinitialization)

QUASAR is a quantum-inspired evolutionary algorithm, efficient for minimizing complex high-dimensional, non-differentiable, and non-parametric objective functions.

• Benefit: Statistically significant improvements in convergence speed and solution quality compared to contemporary optimizers.

• Reference: See experimental trials and analysis: [https://arxiv.org/abs/2511.13843].

Quick Use Example:

import hdim_opt
import numpy as np

# define search space
n_dim = 100
bounds = [(-100,100)] * n_dim

def obj_func(x):
    y = np.sum(x**2)
    return y

# run QUASAR
solution, fitness = hdim_opt.quasar(func=obj_func, bounds=bounds)

HDS Sampler (Hyperellipsoid Density Sampling)

HDS is a non-uniform Quasi-Monte Carlo sampling method, specifically designed to exploit promising regions of the search space.

• Benefit: Provides control over the sample distribution. Results in higher average optimization solution quality when used for population initialization compared to uniform QMC methods.

• Reference: See experimental trials and analysis: [https://arxiv.org/abs/2511.07836].

Quick Use Example:

import hdim_opt

# define search space
n_dim = 2
bounds = [(0,1)] * n_dim

# optional weights
weights = {
            0 : {'center': 0.25, 'std': 0.33},
            1 : {'center': 0.25, 'std': 0.33}
            }

# generate HDS samples
hds_samples = hdim_opt.hds(n_samples=10000, bounds=bounds, 
                    weights=weights, verbose=True)

Additional functions include:

• Sobol sampling (via SciPy):
  • sobol(n_samples, bounds)
• Sensitivity analysis (via SALib):
  • sensitivity(func, bounds)