Dynamic System Optimization

Overview

This repository showcases various optimization techniques applied to dynamic systems, developed within the scope of a university module.
It contains clean, modular implementations of:

• ARX Modeling (AutoRegressive with eXogenous inputs)
• Linear Programming (LP)
• Particle Swarm Optimization (PSO)
• Genetic Algorithm (GA)

Each subproject is self-contained, with input/output examples and mathematical background where necessary.

Key Objectives

• Apply core optimization strategies to real-world dynamic system problems.
• Explore both deterministic and heuristic methods.
• Focus on interpretable code, manual implementations (no scikit-learn for modeling), and scientific reproducibility.

Subprojects

---

1. ARX Modeling

• Implements a manual ARX model without external ML libraries.
• Includes data preprocessing, noise injection, signal smoothing, and model evaluation (RMSE, MAE).
• Input: CSV time series data (u.csv, y.csv).
• Output: Error plots, prediction accuracy.

📁 /ARX-Model

---

2. Linear Programming

• Solves resource allocation and system constraint problems using scipy.optimize.linprog.
• Demonstrates objective function formulation and constraint matrices.

📁 /Linear-Programming

---

3. Particle Swarm Optimization

• Implements a fully custom PSO algorithm from scratch.
• Used to minimize benchmark functions or user-defined loss functions.
• Includes 3D visualization of convergence and parameter tuning interface.

📁 /PSO

---

Technologies Used

• Programming Language: Python
• Libraries: numpy, matplotlib, scipy, pandas, sklearn, mpl_toolkits
• Development Environment: VS Code

Prerequisites

Install the required packages:

pip install numpy pandas matplotlib scipy sklearn mpl_toolkits