OptVisual.jl

Learning Julia right now...

OptVisual.jl is a lightweight package I wrote while learning Julia,
for visualizing simple optimization algorithms.

It provides:

• A set of standard benchmark functions (Sphere, Rosenbrock, Rastrigin, Himmelblau, L1 norm)
• Basic first-order optimization methods (Gradient Descent, Nesterov Accelerated Gradient)
• Visualization of convergence curves and 2D optimization trajectories on contour plots

---

Installation

This package is not yet in the Julia General registry — install directly from GitHub:

using Pkg
Pkg.add(url="https://github.com/LJS42/OptVisual.jl.git")

Example

using OptVisual

# Define a 2D test function (Himmelblau)
fun = Himmelblaufunction(2)

# Define an optimization method (Gradient Descent)
method = GD(0.01, 200, 10)

# Visualize optimization process
optvisual(fun, method; x0=[-3.0, -3.0])

Supports self-defined functions and optimization methods

Function:

struct yourfunction <: AbstractTestFunction
    d::Int #dimension
end

function f(::yourfunction, x::AbstractVector)
    #your function here
end

Method:

struct yourmethod <: AbstractOptMethod
    #your hyperpameter
end

function Opt(fun::AbstractTestFunction, method::yourmethod, x::AbstractVector)
    #your method

    #One step! Example:
    #g = grad(fun, x)
    #x -= method.η * g
    #return x
end