Algorithm optimization

This project is an implementation of several popular optimization algorithms.

Used algorithms

• Grey Wolf Algorithm
• Multi-Swarm Slime Mould Algorithm
• Bat Algorithm
• Butterfly Optimization Algorithm
• Particle Swarm Optimization
• Particle Swarm Optimization with modification from Differential Evolution (mutation and crossover)
• Differential Evolution

Features

• Optimization: Use any of the above-listed algorithms to optimize functions within a given scope.
• Plotting: The program generates plots that track the best solution value for each iteration of the algorithm.
• Comparative Results: You can compare the results of SMA and GWO using custom functions over multiple iterations.
• Function Customization: Users can specify the mathematical function to optimize, along with its scope, dimensionality, and iterations.

Setup and Dependencies

The project uses Python and the following dependencies:

• numpy
• matplotlib
• argparse
• Custom modules for the optimization algorithms (de.py, pso_de.py, ba.py, boa.py, sma.py, gwo.py)

Usage

You can run the script using the command line. The script accepts several arguments to customize the optimization process.

Basic Usage:

python main.py -a [algorithm] -f [function_id]

Where algorithm can be one of the following:

• de (Differential Evolution)
• pso (Particle Swarm Optimization)
• ba (Bat Algorithm)
• boa (Butterfly Optimization Algorithm)
• sma (Slime Mould Algorithm)
• gwo (Grey Wolf Optimizer)

Example:

python main.py -a de -f 1 -ps 50 -d 30 -i 200

Customization Options

You can adjust several parameters depending on the algorithm and function:

• -f, --function: Select which function to optimize (1–6)
• -sf, --scope_flag: Use default function scope
• -s, --scope: Define a custom scope as a tuple (min, max)
• -d, --dimensions: Number of dimensions (default: 20)
• -i, --iterations: Number of iterations (default: 100)
• -f, --function: Select which function to optimize (1–6)
• -sf, --scope_flag: Use default function scope
• -s, --scope: Define a custom scope as a tuple (min, max)
• -d, --dimensions: Number of dimensions (default: 20)
• -i, --iterations: Number of iterations (default: 100)
• -v, --verbose: Whether to print detailed information
• -dw, --differential_weight: Value of differential weight, PSO-DE ONLY (default: 0.5)
• -cp, --crossover_probability: Value of crossover probability, PSO-DE ONLY (default: 0.5)
• -ps, --population_size: Population size, DE/BA/BOA ONLY (default: 50)
• -pn, --particle_number: Number of particles, PSO ONLY (default: 30)
• -iw, --inertia_weight: Weight of inertia, PSO ONLY (default: 0.5)
• -cc, --cognitive_component: Value of cognitive component, PSO ONLY (default: 2)
• -sc, --social_component: Value of social component, PSO ONLY (default: 2)
• -l, --loudness: Initial value of loudness, BA ONLY (default: 0.7)
• -pr, --pulse_rate: Initial value of pulse emission rate, BA ONLY (default: 0.5)
• --alpha: Value of alpha parameter that affects balance between exploration and exploitation, BA ONLY (default: 0.9)
• --gamma: Value of gamma parameter that affects the convergence speed, BA ONLY (default: 0.9)
• --f_min: Minimum frequency value, BA ONLY (default: 0)
• --f_max: Maximum frequency value, BA ONLY (default: 2)
• -sm, --sensory_modality: Value of sensory modality, BOA ONLY (default: 0.4)
• -pe, --power_exponent: Value of power exponent, BOA ONLY (default: 0.8)
• -sp, --switch_probability: Value of switch probability, BOA ONLY (default: 0.8)
• -ss, --swarm_size: Size of swarm, SMA ONLY (default: 20)
• -ns, --num_swarms: Number of swarms, SMA ONLY (default: 10)
• -z, --z: Probability of random updating position, SMA ONLY (default: 0.6)
• -mt, --migration_threshold: Migration threshold, SMA ONLY (default: 0.4)
• -mr, --mutation_rate: Mutation rate, SMA/GWO ONLY (default: 0.4)
• -nil, --no_improve_limit: Limit for result stagnation, SMA/GWO ONLY (default: 30)

License

MIT