DAPAO (Distributed Advanced PArallel Optimization)

A high-performance optimization library that leverages parallel computing for large-scale optimization problems. DAPAO provides efficient implementations of classical and modern optimization algorithms including Steepest Descent, Newton, Modified Newton, BFGS, L-BFGS, and Newton-CG.

Key Features

• Parallel Execution: Utilizes Julia's native parallel computing capabilities for faster convergence on multi-core systems
• Distributed Computing Support: Scales to multiple nodes for handling very large optimization problems
• High Performance: Optimized implementations of gradient-based algorithms
• Flexible API: Simple interface that works with both simple and complex optimization problems
• Extensible Design: Easy to add custom optimization methods

Installation

You can install DAPAO from the Julia REPL using the package manager:

using Pkg
Pkg.add("DAPAO")

Or, in pkg mode (press ] in the REPL):

add DAPAO

Direction Search Methods

• Steepest Descent
• Newton
• Modified Newton
• BFGS
• L-BFGS
• Newton-CG

TO-DO

• Implement BFGS method
• Implement L-BFGS method
• Pass all tests for different functions in test/runtest.jl
• Test scalability

How to use

Basic usage example:

using DAPAO
using LinearAlgebra

# Define your objective function and its gradient
function f(x, p)
    return x[1]^2 + 2x[2]^2
end

function grad(x, p)
    g = zeros(2)
    g[1] = 2x[1]
    g[2] = 4x[2]
    return g
end

# Optimize using Steepest Descent
optfunc = OptimizationFunction(f, grad)
x0 = [-1.2, 1.0]
prob = OptimizationProblem(optfunc, x0)
sol = solve(prob, SteepestDescent())
println("Minimum found at: ", sol.x)
println("Minimum value: ", sol.f_val)

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.